merge pull request #70 from uglytoad/add-images

#55 move support for images to page and add inline images

src/UglyToad.PdfPig.Tests/Graphics/Operations/GraphicsStateOperationTests.cs

@@ -44,7 +44,11 @@
                 }
                 else if (operationType == typeof(EndInlineImage))
                 {
-                    operation = new EndInlineImage(new List<IToken>(), new List<byte>());
+                    operation = new EndInlineImage(new List<byte>());
+                }
+                else if (operationType == typeof(BeginInlineImageData))
+                {
+                    operation = new BeginInlineImageData(new Dictionary<NameToken, IToken>());
                 }
                 else
                 {

src/UglyToad.PdfPig.Tests/Graphics/TestOperationContext.cs

@@ -77,5 +77,17 @@
         public void SetNamedGraphicsState(NameToken stateName)
         {
         }
+
+        public void BeginInlineImage()
+        {
+        }
+
+        public void SetInlineImageProperties(IReadOnlyDictionary<NameToken, IToken> properties)
+        {
+        }
+
+        public void EndInlineImage(IReadOnlyList<byte> bytes)
+        {
+        }
     }
 }

src/UglyToad.PdfPig.Tests/Graphics/TestResourceStore.cs

@@ -1,33 +0,0 @@
-namespace UglyToad.PdfPig.Tests.Graphics
-{
-    using Content;
-    using PdfPig.Fonts;
-    using PdfPig.Tokens;
-
-    internal class TestResourceStore : IResourceStore
-    {
-        public void LoadResourceDictionary(DictionaryToken dictionary, bool isLenientParsing)
-        {
-        }
-
-        public IFont GetFont(NameToken name)
-        {
-            return null;
-        }
-
-        public StreamToken GetXObject(NameToken name)
-        {
-            return null;
-        }
-
-        public DictionaryToken GetExtendedGraphicsStateDictionary(NameToken name)
-        {
-            return null;
-        }
-
-        public IFont GetFontDirectly(IndirectReferenceToken fontReferenceToken, bool isLenientParsing)
-        {
-            return null;
-        }
-    }
-}

src/UglyToad.PdfPig.Tests/Integration/IntegrationDocumentTests.cs

@@ -81,14 +81,14 @@
                 {
                     var page = document.GetPage(i + 1);
 
-                    var images = page.ExperimentalAccess.GetRawImages();
+                    var images = page.GetImages();
 
                     Assert.NotNull(images);
 
                     foreach (var image in images)
                     {
-                        Assert.True(image.Width > 0, $"Image had width of zero on page {i + 1}.");
-                        Assert.True(image.Height > 0, $"Image had height of zero on page {i + 1}.");
+                        Assert.True(image.WidthInSamples > 0, $"Image had width of zero on page {i + 1}.");
+                        Assert.True(image.HeightInSamples > 0, $"Image had height of zero on page {i + 1}.");
                     }
                 }
             }

src/UglyToad.PdfPig.Tests/Integration/OldGutnishTests.cs

@@ -41,7 +41,7 @@
             {
                 var page = document.GetPage(1);
 
-                var images = page.ExperimentalAccess.GetRawImages().ToList();
+                var images = page.GetImages().ToList();
                 Assert.Single(images);
             }
         }

src/UglyToad.PdfPig.Tests/Integration/SinglePageLibreOfficeImages.cs

@@ -0,0 +1,87 @@
+namespace UglyToad.PdfPig.Tests.Integration
+{
+    using System;
+    using System.Linq;
+    using Xunit;
+
+    public class SinglePageLibreOfficeImages
+    {
+        private static string GetFilePath() => IntegrationHelpers.GetDocumentPath(@"Single Page Images - from libre office.pdf");
+
+        [Fact]
+        public void Has3Images()
+        {
+            using (var document = PdfDocument.Open(GetFilePath(), ParsingOptions.LenientParsingOff))
+            {
+                var page = document.GetPage(1);
+                
+                var images = page.GetImages().ToList();
+
+                Assert.Equal(3, images.Count);
+            }
+        }
+
+        [Fact]
+        public void ImagesHaveCorrectDimensionsAndLocations()
+        {
+            using (var document = PdfDocument.Open(GetFilePath(), ParsingOptions.LenientParsingOff))
+            {
+                var page = document.GetPage(1);
+
+                var images = page.GetImages().OrderBy(x => x.Bounds.Width).ToList();
+
+                var pdfPigSquare = images[0];
+
+                Assert.Equal(148.3m, pdfPigSquare.Bounds.Width);
+                Assert.Equal(148.3m, pdfPigSquare.Bounds.Height);
+                Assert.Equal(60.1m, pdfPigSquare.Bounds.Left);
+                Assert.Equal(765.8m, pdfPigSquare.Bounds.Top);
+
+                var pdfPigSquished = images[1];
+
+                Assert.Equal(206.8m, pdfPigSquished.Bounds.Width);
+                Assert.Equal(83.2m, pdfPigSquished.Bounds.Height);
+                Assert.Equal(309.8m, pdfPigSquished.Bounds.Left);
+                Assert.Equal(552.1m, pdfPigSquished.Bounds.Top);
+
+                var birthdayPigs = images[2];
+
+                Assert.Equal(391m, birthdayPigs.Bounds.Width);
+                Assert.Equal(267.1m, birthdayPigs.Bounds.Height);
+                Assert.Equal(102.2m, birthdayPigs.Bounds.Left);
+                Assert.Equal(426.3m, birthdayPigs.Bounds.Top);
+            }
+        }
+
+        [Fact]
+        public void HasCorrectText()
+        {
+            using (var document = PdfDocument.Open(GetFilePath(), ParsingOptions.LenientParsingOff))
+            {
+                var page = document.GetPage(1);
+                Assert.Equal("Oink oink", page.Text);
+            }
+        }
+
+        [Fact]
+        public void CanAccessImageBytesExceptUnsupported()
+        {
+            using (var document = PdfDocument.Open(GetFilePath(), ParsingOptions.LenientParsingOff))
+            {
+                var page = document.GetPage(1);
+                foreach (var image in page.GetImages())
+                {
+                    try
+                    {
+                        Assert.NotNull(image.Bytes);
+                    }
+                    catch (NotSupportedException )
+                    {
+                        // Should allow access to raw bytes.
+                        Assert.NotNull(image.RawBytes);
+                    }
+                }
+            }
+        }
+    }
+}

src/UglyToad.PdfPig.Tests/PublicApiScannerTests.cs

@@ -53,6 +53,8 @@
                 "UglyToad.PdfPig.Annotations.AnnotationType",
                 "UglyToad.PdfPig.Content.Catalog",
                 "UglyToad.PdfPig.Content.DocumentInformation",
+                "UglyToad.PdfPig.Content.InlineImage",
+                "UglyToad.PdfPig.Content.IPdfImage",
                 "UglyToad.PdfPig.Content.Letter",
                 "UglyToad.PdfPig.Content.Page",
                 "UglyToad.PdfPig.Content.PageRotationDegrees",

src/UglyToad.PdfPig.Tests/TestFilterProvider.cs

@@ -11,6 +11,11 @@
             return new List<IFilter>();
         }
 
+        public IReadOnlyList<IFilter> GetNamedFilters(IReadOnlyList<NameToken> names)
+        {
+            return new List<IFilter>();
+        }
+
         public IReadOnlyList<IFilter> GetAllFilters()
         {
             return new List<IFilter>();

src/UglyToad.PdfPig.Tests/Tokens/TestPdfTokenScanner.cs

@@ -42,5 +42,9 @@ namespace UglyToad.PdfPig.Tests.Tokens
         {
             return Objects[reference];
         }
+
+        public void Dispose()
+        {
+        }
     }
 }

src/UglyToad.PdfPig/Content/IPdfImage.cs

@@ -0,0 +1,94 @@
+namespace UglyToad.PdfPig.Content
+{
+    using System.Collections.Generic;
+    using Geometry;
+    using Graphics.Colors;
+    using Graphics.Core;
+    using XObjects;
+
+    /// <summary>
+    /// An image in a PDF document, may be an <see cref="InlineImage"/> or a PostScript image XObject (<see cref="XObjectImage"/>).
+    /// </summary>
+    public interface IPdfImage
+    {
+        /// <summary>
+        /// The placement rectangle of the image in PDF coordinates.
+        /// </summary>
+        PdfRectangle Bounds { get; }
+
+        /// <summary>
+        /// The width of the image in samples.
+        /// </summary>
+        int WidthInSamples { get; }
+
+        /// <summary>
+        /// The height of the image in samples.
+        /// </summary>
+        int HeightInSamples { get; }
+
+        /// <summary>
+        /// The <see cref="ColorSpace"/> used to interpret the image.
+        /// This defines the number of color components per sample, e.g.
+        /// 1 component for <see cref="Graphics.Colors.ColorSpace.DeviceGray"/>,
+        /// 3 components for <see cref="Graphics.Colors.ColorSpace.DeviceRGB"/>,
+        /// 4 components for <see cref="Graphics.Colors.ColorSpace.DeviceCMYK"/>,
+        /// etc.
+        /// This is not defined where <see cref="IsImageMask"/> is <see langword="true"/> and is optional where the image is JPXEncoded for <see cref="XObjectImage"/>.
+        /// </summary>
+        ColorSpace? ColorSpace { get; }
+
+        /// <summary>
+        /// The number of bits used to represent each color component.
+        /// </summary>
+        int BitsPerComponent { get; }
+
+        /// <summary>
+        /// The bytes of the image with any filters decoded.
+        /// If the filter used to encode the bytes is not supported accessing this property will throw, access the <see cref="RawBytes"/>
+        /// instead.
+        /// </summary>
+        IReadOnlyList<byte> Bytes { get; }
+
+        /// <summary>
+        /// The encoded bytes of the image with all filters still applied.
+        /// </summary>
+        IReadOnlyList<byte> RawBytes { get; }
+
+            /// <summary>
+        /// The color rendering intent to be used when rendering the image.
+        /// </summary>
+        RenderingIntent RenderingIntent { get; }
+
+        /// <summary>
+        /// Indicates whether the image is to be treated as an image mask.
+        /// If <see langword="true"/> the image is a monochrome image in which each sample
+        /// is specified by a single bit (<see cref="BitsPerComponent"/> is 1).
+        /// The image represents a stencil where sample values represent places on the page
+        /// that should be marked with the current color or masked (not marked).
+        /// </summary>
+        bool IsImageMask { get; }
+
+        /// <summary>
+        /// Describes how to map image samples into the values appropriate for the
+        /// <see cref="ColorSpace"/>.
+        /// The image data is initially composed of values in the range 0 to 2^n - 1
+        /// where n is <see cref="BitsPerComponent"/>.
+        /// The decode array contains a pair of numbers for each component in the <see cref="ColorSpace"/>.
+        /// The value from the image data is then interpolated into the values relevant to the <see cref="ColorSpace"/>
+        /// using the corresponding values of the decode array.
+        /// </summary>
+        IReadOnlyList<decimal> Decode { get; }
+
+        /// <summary>
+        /// Specifies whether interpolation is to be performed. Interpolation smooths images where a single component in the image
+        /// as defined may correspond to many pixels on the output device. The interpolation algorithm is implementation
+        /// dependent and is not defined by the specification.
+        /// </summary>
+        bool Interpolate { get; }
+
+        /// <summary>
+        /// Whether this image is an <see cref="InlineImage"/> or a <see cref="XObjectImage"/>.
+        /// </summary>
+        bool IsInlineImage { get; }
+    }
+}

src/UglyToad.PdfPig/Content/IResourceStore.cs

@@ -14,5 +14,7 @@
         DictionaryToken GetExtendedGraphicsStateDictionary(NameToken name);
 
         IFont GetFontDirectly(IndirectReferenceToken fontReferenceToken, bool isLenientParsing);
+
+        bool TryGetNamedColorSpace(NameToken name, out IToken namedColorSpace);
     }
 }

src/UglyToad.PdfPig/Content/InlineImage.cs

@@ -0,0 +1,98 @@
+namespace UglyToad.PdfPig.Content
+{
+    using System;
+    using System.Collections.Generic;
+    using System.Linq;
+    using Filters;
+    using Geometry;
+    using Graphics.Colors;
+    using Graphics.Core;
+    using Tokens;
+
+    /// <inheritdoc />
+    /// <summary>
+    /// A small image that is completely defined directly inline within a <see cref="T:UglyToad.PdfPig.Content.Page" />'s content stream.
+    /// </summary>
+    public class InlineImage : IPdfImage
+    {
+        private readonly Lazy<IReadOnlyList<byte>> bytesFactory;
+
+        /// <inheritdoc />
+        public PdfRectangle Bounds { get; }
+
+        /// <inheritdoc />
+        public int WidthInSamples { get; }
+
+        /// <inheritdoc />
+        public int HeightInSamples { get; }
+
+        /// <inheritdoc />
+        public ColorSpace? ColorSpace { get; }
+
+        /// <inheritdoc />
+        public int BitsPerComponent { get; }
+
+        /// <inheritdoc />
+        public bool IsImageMask { get; }
+
+        /// <inheritdoc />
+        public IReadOnlyList<decimal> Decode { get; }
+
+        /// <inheritdoc />
+        public bool IsInlineImage { get; } = true;
+
+        /// <inheritdoc />
+        public RenderingIntent RenderingIntent { get; }
+
+        /// <inheritdoc />
+        public bool Interpolate { get; }
+
+        /// <inheritdoc />
+        public IReadOnlyList<byte> Bytes => bytesFactory.Value;
+
+        /// <inheritdoc />
+        public IReadOnlyList<byte> RawBytes { get; }
+
+        /// <summary>
+        /// Create a new <see cref="InlineImage"/>.
+        /// </summary>
+        internal InlineImage(PdfRectangle bounds, int widthInSamples, int heightInSamples, int bitsPerComponent, bool isImageMask,
+            RenderingIntent renderingIntent,
+            bool interpolate,
+            ColorSpace? colorSpace,
+            IReadOnlyList<decimal> decode,
+            IReadOnlyList<byte> bytes,
+            IReadOnlyList<IFilter> filters,
+            DictionaryToken streamDictionary)
+        {
+            Bounds = bounds;
+            WidthInSamples = widthInSamples;
+            HeightInSamples = heightInSamples;
+            ColorSpace = colorSpace;
+            Decode = decode;
+            BitsPerComponent = bitsPerComponent;
+            IsImageMask = isImageMask;
+            RenderingIntent = renderingIntent;
+            Interpolate = interpolate;
+
+            RawBytes = bytes;
+            bytesFactory = new Lazy<IReadOnlyList<byte>>(() =>
+            {
+                var b = bytes.ToArray();
+                for (var i = 0; i < filters.Count; i++)
+                {
+                    var filter = filters[i];
+                    b = filter.Decode(b, streamDictionary, i);
+                }
+
+                return b;
+            });
+        }
+
+        /// <inheritdoc />
+        public override string ToString()
+        {
+            return $"Inline Image (w {Bounds.Width}, h {Bounds.Height})";
+        }
+    }
+}

src/UglyToad.PdfPig/Content/Page.cs

@@ -8,7 +8,6 @@
     using Tokens;
     using Util;
     using Util.JetBrains.Annotations;
-    using XObjects;
     using Geometry;
 
     /// <summary>
@@ -60,7 +59,7 @@
         public decimal Height { get; }
 
         /// <summary>
-        /// The size of the page according to the standard page sizes or Custom if no matching standard size found.
+        /// The size of the page according to the standard page sizes or <see cref="PageSize.Custom"/> if no matching standard size found.
         /// </summary>
         public PageSize Size { get; }
 
@@ -131,6 +130,11 @@
             return (wordExtractor ?? DefaultWordExtractor.Instance).GetWords(Letters);
         }
 
+        /// <summary>
+        /// Gets any images on the page.
+        /// </summary>
+        public IEnumerable<IPdfImage> GetImages() => Content.GetImages();
+
         /// <summary>
         /// Provides access to useful members which will change in future releases.
         /// </summary>
@@ -150,16 +154,6 @@
                 this.annotationProvider = annotationProvider;
             }
 
-            /// <summary>
-            /// Retrieve any images referenced in this page's content.
-            /// These are returned as <see cref="XObjectImage"/>s which are 
-            /// raw data from the PDF's content rather than images.
-            /// </summary>
-            public IEnumerable<XObjectImage> GetRawImages()
-            {
-                return page.Content.GetImages();
-            }
-
             /// <summary>
             /// Get the annotation objects from the page.
             /// </summary>

src/UglyToad.PdfPig/Content/PageContent.cs

@@ -1,14 +1,17 @@
 namespace UglyToad.PdfPig.Content
 {
+    using System;
     using System.Collections.Generic;
+    using Filters;
     using Graphics;
     using Graphics.Operations;
     using Tokenization.Scanner;
     using XObjects;
-    using UglyToad.PdfPig.Geometry;
+    using Geometry;
+    using Util;
 
     /// <summary>
-    /// 
+    /// Wraps content parsed from a page content stream for access.
     /// </summary>
     /// <remarks>
     /// This should contain a replayable stack of drawing instructions for page content
@@ -16,36 +19,46 @@
     /// </remarks>
     internal class PageContent
     {
-        private readonly IReadOnlyDictionary<XObjectType, List<XObjectContentRecord>> xObjects;
+        private readonly IReadOnlyList<Union<XObjectContentRecord, InlineImage>> images;
         private readonly IPdfTokenScanner pdfScanner;
-        private readonly XObjectFactory xObjectFactory;
+        private readonly IFilterProvider filterProvider;
+        private readonly IResourceStore resourceStore;
         private readonly bool isLenientParsing;
 
         internal IReadOnlyList<IGraphicsStateOperation> GraphicsStateOperations { get; }
 
         public IReadOnlyList<Letter> Letters { get; }
+
         public IReadOnlyList<PdfPath> Paths { get; }
 
-        internal PageContent(IReadOnlyList<IGraphicsStateOperation> graphicsStateOperations, IReadOnlyList<Letter> letters, List<PdfPath> paths,
-            IReadOnlyDictionary<XObjectType, List<XObjectContentRecord>> xObjects,
+        internal PageContent(IReadOnlyList<IGraphicsStateOperation> graphicsStateOperations, IReadOnlyList<Letter> letters,
+            IReadOnlyList<PdfPath> paths,
+            IReadOnlyList<Union<XObjectContentRecord, InlineImage>> images,
             IPdfTokenScanner pdfScanner,
-            XObjectFactory xObjectFactory,
+            IFilterProvider filterProvider,
+            IResourceStore resourceStore,
             bool isLenientParsing)
         {
             GraphicsStateOperations = graphicsStateOperations;
             Letters = letters;
             Paths = paths;
-            this.xObjects = xObjects;
-            this.pdfScanner = pdfScanner;
-            this.xObjectFactory = xObjectFactory;
+            this.images = images;
+            this.pdfScanner = pdfScanner ?? throw new ArgumentNullException(nameof(pdfScanner));
+            this.filterProvider = filterProvider ?? throw new ArgumentNullException(nameof(filterProvider));
+            this.resourceStore = resourceStore ?? throw new ArgumentNullException(nameof(resourceStore));
             this.isLenientParsing = isLenientParsing;
         }
 
-        public IEnumerable<XObjectImage> GetImages()
+        public IEnumerable<IPdfImage> GetImages()
         {
-            foreach (var contentRecord in xObjects[XObjectType.Image])
+            foreach (var image in images)
             {
-                yield return xObjectFactory.CreateImage(contentRecord, pdfScanner, isLenientParsing);
+
+                IPdfImage result = null;
+                image.Match(x => { result = XObjectFactory.ReadImage(x, pdfScanner, filterProvider, resourceStore, isLenientParsing); },
+                    x => { result = x; });
+
+                yield return result;
             }
         }
     }

src/UglyToad.PdfPig/Content/ResourceStore.cs

@@ -8,7 +8,7 @@
     using Tokenization.Scanner;
     using Tokens;
 
-    internal class ResourceContainer : IResourceStore
+    internal class ResourceStore : IResourceStore
     {
         private readonly IPdfTokenScanner scanner;
         private readonly IFontFactory fontFactory;
@@ -18,7 +18,9 @@
 
         private readonly Dictionary<NameToken, DictionaryToken> extendedGraphicsStates = new Dictionary<NameToken, DictionaryToken>();
 
-        public ResourceContainer(IPdfTokenScanner scanner, IFontFactory fontFactory)
+        private readonly Dictionary<NameToken, NameToken> colorSpaceNames = new Dictionary<NameToken, NameToken>();
+
+        public ResourceStore(IPdfTokenScanner scanner, IFontFactory fontFactory)
         {
             this.scanner = scanner;
             this.fontFactory = fontFactory;
@@ -58,6 +60,39 @@
                     extendedGraphicsStates[name] = state;
                 }
             }
+
+            if (resourceDictionary.TryGet(NameToken.ColorSpace, scanner, out DictionaryToken colorSpaceDictionary))
+            {
+                foreach (var nameColorSpacePair in colorSpaceDictionary.Data)
+                {
+                    var name = NameToken.Create(nameColorSpacePair.Key);
+
+                    if (DirectObjectFinder.TryGet(nameColorSpacePair.Value, scanner, out NameToken colorSpaceName))
+                    {
+                        colorSpaceNames[name] = colorSpaceName;
+                    }
+                    else if (DirectObjectFinder.TryGet(nameColorSpacePair.Value, scanner, out ArrayToken colorSpaceArray))
+                    {
+                        if (colorSpaceArray.Length == 0)
+                        {
+                            throw new PdfDocumentFormatException($"Empty ColorSpace array encountered in page resource dictionary: {resourceDictionary}.");
+                        }
+
+                        var first = colorSpaceArray.Data[0];
+
+                        if (!(first is NameToken arrayNamedColorSpace))
+                        {
+                            throw new PdfDocumentFormatException($"Invalid ColorSpace array encountered in page resource dictionary: {colorSpaceArray}.");
+                        }
+
+                        colorSpaceNames[name] = arrayNamedColorSpace;
+                    }
+                    else
+                    {
+                        throw new PdfDocumentFormatException($"Invalid ColorSpace token encountered in page resource dictionary: {nameColorSpacePair.Value}.");
+                    }
+                }
+            }
         }
 
         private void LoadFontDictionary(DictionaryToken fontDictionary, bool isLenientParsing)
@@ -115,6 +150,25 @@
             return font;
         }
 
+        public bool TryGetNamedColorSpace(NameToken name, out IToken namedToken)
+        {
+            namedToken = null;
+
+            if (name == null)
+            {
+                throw new ArgumentNullException(nameof(name));
+            }
+
+            if (!colorSpaceNames.TryGetValue(name, out var colorSpaceName))
+            {
+                return false;
+            }
+
+            namedToken = colorSpaceName;
+
+            return true;
+        }
+
         public StreamToken GetXObject(NameToken name)
         {
             var reference = currentResourceState[name];

src/UglyToad.PdfPig/Filters/Ascii85Filter.cs

@@ -33,8 +33,6 @@
             using (var stream = new MemoryStream())
             using (var writer = new BinaryWriter(stream))
             {
-
-
                 for (var i = 0; i < input.Count; i++)
                 {
                     var value = input[i];

src/UglyToad.PdfPig/Filters/CcittFaxDecodeFilter.cs

@@ -0,0 +1,15 @@
+namespace UglyToad.PdfPig.Filters
+{
+    using System;
+    using System.Collections.Generic;
+    using Tokens;
+
+    internal class CcittFaxDecodeFilter : IFilter
+    {
+        public byte[] Decode(IReadOnlyList<byte> input, DictionaryToken streamDictionary, int filterIndex)
+        {
+            throw new NotSupportedException("The CCITT Fax Filter for image data is not currently supported. " +
+                                            "Try accessing the raw compressed data directly.");
+        }
+    }
+}

src/UglyToad.PdfPig/Filters/DctDecodeFilter.cs

@@ -8,7 +8,8 @@
     {
         public byte[] Decode(IReadOnlyList<byte> input, DictionaryToken streamDictionary, int filterIndex)
         {
-            throw new NotImplementedException();
+            throw new NotSupportedException("The DST (Discrete Cosine Transform) Filter indicates data is encoded in JPEG format. " +
+                                            "This filter is not currently supported but the raw data can be supplied to JPEG supporting libraries.");
         }
     }
 }

src/UglyToad.PdfPig/Filters/IFilterProvider.cs

@@ -7,6 +7,8 @@
     {
         IReadOnlyList<IFilter> GetFilters(DictionaryToken dictionary);
 
+        IReadOnlyList<IFilter> GetNamedFilters(IReadOnlyList<NameToken> names);
+
         IReadOnlyList<IFilter> GetAllFilters();
     }
 }

src/UglyToad.PdfPig/Filters/Jbig2DecodeFilter.cs

@@ -0,0 +1,15 @@
+namespace UglyToad.PdfPig.Filters
+{
+    using System;
+    using System.Collections.Generic;
+    using Tokens;
+
+    internal class Jbig2DecodeFilter : IFilter
+    {
+        public byte[] Decode(IReadOnlyList<byte> input, DictionaryToken streamDictionary, int filterIndex)
+        {
+            throw new NotSupportedException("The JBIG2 Filter for monochrome image data is not currently supported. " +
+                                            "Try accessing the raw compressed data directly.");
+        }
+    }
+}

src/UglyToad.PdfPig/Filters/JpxDecodeFilter.cs

@@ -0,0 +1,15 @@
+namespace UglyToad.PdfPig.Filters
+{
+    using System;
+    using System.Collections.Generic;
+    using Tokens;
+
+    internal class JpxDecodeFilter : IFilter
+    {
+        public byte[] Decode(IReadOnlyList<byte> input, DictionaryToken streamDictionary, int filterIndex)
+        {
+            throw new NotSupportedException("The JPX Filter (JPEG2000) for image data is not currently supported. " +
+                                            "Try accessing the raw compressed data directly.");
+        }
+    }
+}

src/UglyToad.PdfPig/Filters/MemoryFilterProvider.cs

@@ -2,6 +2,7 @@
 {
     using System;
     using System.Collections.Generic;
+    using System.Linq;
     using Exceptions;
     using Logging;
     using Tokens;
@@ -15,7 +16,11 @@
         {
             var ascii85 = new Ascii85Filter();
             var asciiHex = new AsciiHexDecodeFilter();
+            var ccitt = new CcittFaxDecodeFilter();
+            var dct = new DctDecodeFilter();
             var flate = new FlateFilter(decodeParameterResolver, pngPredictor, log);
+            var jbig2 = new Jbig2DecodeFilter();
+            var jpx = new JpxDecodeFilter();
             var runLength = new RunLengthFilter();
             var lzw = new LzwFilter(decodeParameterResolver, pngPredictor);
 
@@ -25,8 +30,14 @@
                 {NameToken.Ascii85DecodeAbbreviation.Data, ascii85},
                 {NameToken.AsciiHexDecode.Data, asciiHex},
                 {NameToken.AsciiHexDecodeAbbreviation.Data, asciiHex},
+                {NameToken.CcittfaxDecode.Data, ccitt},
+                {NameToken.CcittfaxDecodeAbbreviation.Data, ccitt},
+                {NameToken.DctDecode.Data, dct},
+                {NameToken.DctDecodeAbbreviation.Data, dct},
                 {NameToken.FlateDecode.Data, flate},
                 {NameToken.FlateDecodeAbbreviation.Data, flate},
+                {NameToken.Jbig2Decode.Data, jbig2},
+                {NameToken.JpxDecode.Data, jpx},
                 {NameToken.RunLengthDecode.Data, runLength},
                 {NameToken.RunLengthDecodeAbbreviation.Data, runLength},
                 {NameToken.LzwDecode, lzw},
@@ -65,6 +76,23 @@
             }
         }
 
+        public IReadOnlyList<IFilter> GetNamedFilters(IReadOnlyList<NameToken> names)
+        {
+            if (names == null)
+            {
+                throw new ArgumentNullException(nameof(names));
+            }
+
+            var result = new List<IFilter>();
+
+            foreach (var name in names)
+            {
+                result.Add(GetFilterStrict(name));
+            }
+
+            return result;
+        }
+        
         private IFilter GetFilterStrict(string name)
         {
             if (!filterInstances.TryGetValue(name, out var factory))
@@ -77,7 +105,7 @@
 
         public IReadOnlyList<IFilter> GetAllFilters()
         {
-            throw new System.NotImplementedException();
+            return filterInstances.Values.Distinct().ToList();
         }
     }
 }

src/UglyToad.PdfPig/Graphics/ContentStreamProcessor.cs

@@ -6,6 +6,8 @@
     using Colors;
     using Content;
     using Core;
+    using Exceptions;
+    using Filters;
     using Fonts;
     using Geometry;
     using IO;
@@ -19,27 +21,43 @@
 
     internal class ContentStreamProcessor : IOperationContext
     {
+        /// <summary>
+        /// Stores each letter as it is encountered in the content stream.
+        /// </summary>
+        private readonly List<Letter> letters = new List<Letter>();
+
+        /// <summary>
+        /// Stores each path as it is encountered in the content stream.
+        /// </summary>
         private readonly List<PdfPath> paths = new List<PdfPath>();
+
+        /// <summary>
+        /// Stores a link to each image (either inline or XObject) as it is encountered in the content stream.
+        /// </summary>
+        private readonly List<Union<XObjectContentRecord, InlineImage>> images = new List<Union<XObjectContentRecord, InlineImage>>();
+
         private readonly IResourceStore resourceStore;
         private readonly UserSpaceUnit userSpaceUnit;
         private readonly PageRotationDegrees rotation;
         private readonly bool isLenientParsing;
         private readonly IPdfTokenScanner pdfScanner;
-        private readonly XObjectFactory xObjectFactory;
+        private readonly IFilterProvider filterProvider;
         private readonly ILog log;
 
         private Stack<CurrentGraphicsState> graphicsStack = new Stack<CurrentGraphicsState>();
-        private IFont activeExtendedGraphicsStateFont = null;
+        private IFont activeExtendedGraphicsStateFont;
+        private InlineImageBuilder inlineImageBuilder;
 
-        //a sequence number of ShowText operation to determine whether letters belong to same operation or not (letters that belong to different operations have less changes to belong to same word)
-        private int textSequence = 0;
+        /// <summary>
+        /// A counter to track individual calls to <see cref="ShowText"/> operations used to determine if letters are likely to be
+        /// in the same word/group. This exposes internal grouping of letters used by the PDF creator which may correspond to the
+        /// intended grouping of letters into words.
+        /// </summary>
+        private int textSequence;
 
         public TextMatrices TextMatrices { get; } = new TextMatrices();
 
-        public TransformationMatrix CurrentTransformationMatrix
-        {
-            get { return GetCurrentState().CurrentTransformationMatrix; }
-        }
+        public TransformationMatrix CurrentTransformationMatrix => GetCurrentState().CurrentTransformationMatrix;
 
         public PdfPath CurrentPath { get; private set; }
 
@@ -56,18 +74,18 @@
             {XObjectType.PostScript, new List<XObjectContentRecord>()}
         };
 
-        public List<Letter> Letters = new List<Letter>();
-        public ContentStreamProcessor(PdfRectangle cropBox, IResourceStore resourceStore, UserSpaceUnit userSpaceUnit, PageRotationDegrees rotation, bool isLenientParsing,
+        public ContentStreamProcessor(PdfRectangle cropBox, IResourceStore resourceStore, UserSpaceUnit userSpaceUnit, PageRotationDegrees rotation,
+            bool isLenientParsing,
             IPdfTokenScanner pdfScanner,
-            XObjectFactory xObjectFactory,
+            IFilterProvider filterProvider,
             ILog log)
         {
             this.resourceStore = resourceStore;
             this.userSpaceUnit = userSpaceUnit;
             this.rotation = rotation;
             this.isLenientParsing = isLenientParsing;
-            this.pdfScanner = pdfScanner;
-            this.xObjectFactory = xObjectFactory;
+            this.pdfScanner = pdfScanner ?? throw new ArgumentNullException(nameof(pdfScanner));
+            this.filterProvider = filterProvider ?? throw new ArgumentNullException(nameof(filterProvider));
             this.log = log;
             graphicsStack.Push(new CurrentGraphicsState());
             ColorSpaceContext = new ColorSpaceContext(GetCurrentState);
@@ -75,11 +93,11 @@
 
         public PageContent Process(IReadOnlyList<IGraphicsStateOperation> operations)
         {
-            var currentState = CloneAllStates();
+            CloneAllStates();
 
             ProcessOperations(operations);
 
-            return new PageContent(operations, Letters, paths, xObjects, pdfScanner, xObjectFactory, isLenientParsing);
+            return new PageContent(operations, letters, paths, images, pdfScanner, filterProvider, resourceStore, isLenientParsing);
         }
 
         private void ProcessOperations(IReadOnlyList<IGraphicsStateOperation> operations)
@@ -265,7 +283,7 @@
             var xObjectStream = resourceStore.GetXObject(xObjectName);
 
             // For now we will determine the type and store the object with the graphics state information preceding it.
-            // Then consumers of the page can request the object/s to be retrieved by type.
+            // Then consumers of the page can request the object(s) to be retrieved by type.
             var subType = (NameToken)xObjectStream.StreamDictionary.Data[NameToken.Subtype.Data];
 
             var state = GetCurrentState();
@@ -274,15 +292,15 @@
 
             if (subType.Equals(NameToken.Ps))
             {
-                xObjects[XObjectType.PostScript].Add(new XObjectContentRecord(XObjectType.PostScript, xObjectStream, matrix));
+                xObjects[XObjectType.PostScript].Add(new XObjectContentRecord(XObjectType.PostScript, xObjectStream, matrix, state.RenderingIntent));
             }
             else if (subType.Equals(NameToken.Image))
             {
-                xObjects[XObjectType.Image].Add(new XObjectContentRecord(XObjectType.Image, xObjectStream, matrix));
+                images.Add(Union<XObjectContentRecord, InlineImage>.One(new XObjectContentRecord(XObjectType.Image, xObjectStream, matrix, state.RenderingIntent)));
             }
             else if (subType.Equals(NameToken.Form))
             {
-                xObjects[XObjectType.Form].Add(new XObjectContentRecord(XObjectType.Form, xObjectStream, matrix));
+                xObjects[XObjectType.Form].Add(new XObjectContentRecord(XObjectType.Form, xObjectStream, matrix, state.RenderingIntent));
             }
             else
             {
@@ -361,6 +379,52 @@
             }
         }
 
+        public void BeginInlineImage()
+        {
+            if (inlineImageBuilder != null && !isLenientParsing)
+            {
+                throw new PdfDocumentFormatException("Begin inline image (BI) command encountered while another inline image was active.");
+            }
+
+            inlineImageBuilder = new InlineImageBuilder();
+        }
+
+        public void SetInlineImageProperties(IReadOnlyDictionary<NameToken, IToken> properties)
+        {
+            if (inlineImageBuilder == null)
+            {
+                if (isLenientParsing)
+                {
+                    return;
+                }
+
+                throw new PdfDocumentFormatException("Begin inline image data (ID) command encountered without a corresponding begin inline image (BI) command.");
+            }
+
+            inlineImageBuilder.Properties = properties;
+        }
+
+        public void EndInlineImage(IReadOnlyList<byte> bytes)
+        {
+            if (inlineImageBuilder == null)
+            {
+                if (isLenientParsing)
+                {
+                    return;
+                }
+
+                throw new PdfDocumentFormatException("End inline image (EI) command encountered without a corresponding begin inline image (BI) command.");
+            }
+
+            inlineImageBuilder.Bytes = bytes;
+
+            var image = inlineImageBuilder.CreateInlineImage(CurrentTransformationMatrix, filterProvider, pdfScanner, GetCurrentState().RenderingIntent, resourceStore);
+
+            images.Add(Union<XObjectContentRecord, InlineImage>.Two(image));
+
+            inlineImageBuilder = null;
+        }
+
         private void AdjustTextMatrix(decimal tx, decimal ty)
         {
             var matrix = TransformationMatrix.GetTranslationMatrix(tx, ty);
@@ -390,7 +454,7 @@
                 pointSize,
                 textSequence);
 
-            Letters.Add(letter);
+            letters.Add(letter);
         }
     }
 }

src/UglyToad.PdfPig/Graphics/IOperationContext.cs

@@ -4,7 +4,7 @@
     using Geometry;
     using IO;
     using Tokens;
-    using UglyToad.PdfPig.Core;
+    using PdfPig.Core;
     using Util.JetBrains.Annotations;
 
     /// <summary>
@@ -104,5 +104,20 @@
         /// </summary>
         /// <param name="stateName">The name of the state to apply.</param>
         void SetNamedGraphicsState(NameToken stateName);
+
+        /// <summary>
+        /// Indicate that an inline image is being defined.
+        /// </summary>
+        void BeginInlineImage();
+
+        /// <summary>
+        /// Define the properties of the inline image currently being drawn.
+        /// </summary>
+        void SetInlineImageProperties(IReadOnlyDictionary<NameToken, IToken> properties);
+
+        /// <summary>
+        /// Indicates that the current inline image is complete.
+        /// </summary>
+        void EndInlineImage(IReadOnlyList<byte> bytes);
     }
 }

src/UglyToad.PdfPig/Graphics/InlineImageBuilder.cs

@@ -0,0 +1,213 @@
+namespace UglyToad.PdfPig.Graphics
+{
+    using System;
+    using System.Collections.Generic;
+    using System.Linq;
+    using Colors;
+    using Content;
+    using Core;
+    using Exceptions;
+    using Filters;
+    using Geometry;
+    using PdfPig.Core;
+    using Tokenization.Scanner;
+    using Tokens;
+    using Util;
+
+    internal class InlineImageBuilder
+    {
+        public IReadOnlyDictionary<NameToken, IToken> Properties { get; set; }
+
+        public IReadOnlyList<byte> Bytes { get; set; }
+
+        public InlineImage CreateInlineImage(TransformationMatrix transformationMatrix, IFilterProvider filterProvider,
+            IPdfTokenScanner tokenScanner,
+            RenderingIntent defaultRenderingIntent,
+            IResourceStore resourceStore)
+        {
+            if (Properties == null || Bytes == null)
+            {
+                throw new InvalidOperationException($"Inline image builder not completely defined before calling {nameof(CreateInlineImage)}.");
+            }
+
+            bool TryMapColorSpace(NameToken name, out ColorSpace colorSpaceResult)
+            {
+                if (name.TryMapToColorSpace(out colorSpaceResult))
+                {
+                    return true;
+                }
+
+                if (TryExtendedColorSpaceNameMapping(name, out colorSpaceResult))
+                {
+                    return true;
+                }
+
+                if (!resourceStore.TryGetNamedColorSpace(name, out var colorSpaceNamedToken) || !(colorSpaceNamedToken is NameToken newName))
+                {
+                    return false;
+                }
+
+                if (newName.TryMapToColorSpace(out colorSpaceResult))
+                {
+                    return true;
+                }
+
+                if (TryExtendedColorSpaceNameMapping(newName, out colorSpaceResult))
+                {
+                    return true;
+                }
+
+                return false;
+            }
+            
+            var bounds = transformationMatrix.Transform(new PdfRectangle(new PdfPoint(1, 1),
+                new PdfPoint(0, 0)));
+
+            var width = GetByKeys<NumericToken>(NameToken.Width, NameToken.W, true).Int;
+
+            var height = GetByKeys<NumericToken>(NameToken.Height, NameToken.H, true).Int;
+
+            var maskToken = GetByKeys<BooleanToken>(NameToken.ImageMask, NameToken.Im, false);
+
+            var isMask = maskToken?.Data == true;
+
+            var bitsPerComponent = GetByKeys<NumericToken>(NameToken.BitsPerComponent, NameToken.Bpc, !isMask)?.Int ?? 1;
+
+            var colorSpace = default(ColorSpace?);
+
+            if (!isMask)
+            {
+                var colorSpaceName = GetByKeys<NameToken>(NameToken.ColorSpace, NameToken.Cs, false);
+
+                if (colorSpaceName == null)
+                {
+                    var colorSpaceArray = GetByKeys<ArrayToken>(NameToken.ColorSpace, NameToken.Cs, true);
+
+                    if (colorSpaceArray.Length == 0)
+                    {
+                        throw new PdfDocumentFormatException("Empty ColorSpace array defined for inline image.");
+                    }
+
+                    if (!(colorSpaceArray.Data[0] is NameToken firstColorSpaceName))
+                    {
+                        throw new PdfDocumentFormatException($"Invalid ColorSpace array defined for inline image: {colorSpaceArray}.");
+                    }
+
+                    if (!TryMapColorSpace(firstColorSpaceName, out var colorSpaceMapped))
+                    {
+                        throw new PdfDocumentFormatException($"Invalid ColorSpace defined for inline image: {firstColorSpaceName}.");
+                    }
+
+                    colorSpace = colorSpaceMapped;
+                }
+                else
+                {
+                    if (!TryMapColorSpace(colorSpaceName, out var colorSpaceMapped))
+                    {
+                        throw new PdfDocumentFormatException($"Invalid ColorSpace defined for inline image: {colorSpaceName}.");
+                    }
+
+                    colorSpace = colorSpaceMapped;
+                }
+            }
+
+            var renderingIntent = GetByKeys<NameToken>(NameToken.Intent, null, false)?.Data?.ToRenderingIntent() ?? defaultRenderingIntent;
+
+            var filterNames = new List<NameToken>();
+
+            var filterName = GetByKeys<NameToken>(NameToken.Filter, NameToken.F, false);
+
+            if (filterName == null)
+            {
+                var filterArray = GetByKeys<ArrayToken>(NameToken.Filter, NameToken.F, false);
+
+                if (filterArray != null)
+                {
+                    filterNames.AddRange(filterArray.Data.OfType<NameToken>());
+                }
+            }
+            else
+            {
+                filterNames.Add(filterName);
+            }
+
+            var filters = filterProvider.GetNamedFilters(filterNames);
+
+            var decodeRaw = GetByKeys<ArrayToken>(NameToken.Decode, NameToken.D, false) ?? new ArrayToken(EmptyArray<IToken>.Instance);
+
+            var decode = decodeRaw.Data.OfType<NumericToken>().Select(x => x.Data).ToArray();
+
+            var filterDictionaryEntries = new Dictionary<NameToken, IToken>();
+            var decodeParamsDict = GetByKeys<DictionaryToken>(NameToken.DecodeParms, NameToken.Dp, false);
+
+            if (decodeParamsDict == null)
+            {
+                var decodeParamsArray = GetByKeys<ArrayToken>(NameToken.DecodeParms, NameToken.Dp, false);
+
+                if (decodeParamsArray != null)
+                {
+                    filterDictionaryEntries[NameToken.DecodeParms] = decodeParamsArray;
+                }
+            }
+            else
+            {
+                filterDictionaryEntries[NameToken.DecodeParms] = decodeParamsDict;
+            }
+
+            var streamDictionary = new DictionaryToken(filterDictionaryEntries);
+
+            var interpolate = GetByKeys<BooleanToken>(NameToken.Interpolate, NameToken.I, false)?.Data ?? false;
+
+            return new InlineImage(bounds, width, height, bitsPerComponent, isMask, renderingIntent, interpolate, colorSpace, decode, Bytes,
+                filters,
+                streamDictionary);
+        }
+
+        private static bool TryExtendedColorSpaceNameMapping(NameToken name, out ColorSpace result)
+        {
+            result = ColorSpace.DeviceGray;
+
+            switch (name.Data)
+            {
+                case "G":
+                    result = ColorSpace.DeviceGray;
+                    return true;
+                case "RGB":
+                    result = ColorSpace.DeviceRGB;
+                    return true;
+                case "CMYK":
+                    result = ColorSpace.DeviceCMYK;
+                    return true;
+                case "I":
+                    result = ColorSpace.Indexed;
+                    return true;
+            }
+
+            return false;
+        }
+
+        // ReSharper disable once ParameterOnlyUsedForPreconditionCheck.Local
+        private T GetByKeys<T>(NameToken name1, NameToken name2, bool required) where T : IToken
+        {
+            if (Properties.TryGetValue(name1, out var val) && val is T result)
+            {
+                return result;
+            }
+
+            if (name2 != null)
+            {
+                if (Properties.TryGetValue(name2, out val) && val is T result2)
+                {
+                    return result2;
+                }
+            }
+
+            if (required)
+            {
+                throw new PdfDocumentFormatException($"Inline image dictionary missing required entry {name1}/{name2}.");
+            }
+
+            return default(T);
+        }
+    }
+}

src/UglyToad.PdfPig/Graphics/Operations/InlineImages/BeginInlineImage.cs

@@ -28,6 +28,7 @@
         /// <inheritdoc />
         public void Run(IOperationContext operationContext)
         {
+            operationContext.BeginInlineImage();
         }
 
         /// <inheritdoc />

src/UglyToad.PdfPig/Graphics/Operations/InlineImages/BeginInlineImageData.cs

@@ -1,6 +1,9 @@
 namespace UglyToad.PdfPig.Graphics.Operations.InlineImages
 {
+    using System;
+    using System.Collections.Generic;
     using System.IO;
+    using Tokens;
 
     /// <inheritdoc />
     /// <summary>
@@ -13,21 +16,26 @@
         /// </summary>
         public const string Symbol = "ID";
         
-        /// <summary>
-        /// The instance of the <see cref="BeginInlineImageData"/> operation.
-        /// </summary>
-        public static readonly BeginInlineImageData Value = new BeginInlineImageData();
-
         /// <inheritdoc />
         public string Operator => Symbol;
 
-        private BeginInlineImageData()
+        /// <summary>
+        /// The key-value pairs which specify attributes of the following image.
+        /// </summary>
+        public IReadOnlyDictionary<NameToken, IToken> Dictionary { get; }
+
+        /// <summary>
+        /// Create a new <see cref="BeginInlineImageData"/>.
+        /// </summary>
+        public BeginInlineImageData(IReadOnlyDictionary<NameToken, IToken> dictionary)
         {
+            Dictionary = dictionary ?? throw new ArgumentNullException(nameof(dictionary));
         }
 
         /// <inheritdoc />
         public void Run(IOperationContext operationContext)
         {
+            operationContext.SetInlineImageProperties(Dictionary);
         }
 
         /// <inheritdoc />

src/UglyToad.PdfPig/Graphics/Operations/InlineImages/EndInlineImage.cs

@@ -3,7 +3,6 @@
     using System;
     using System.Collections.Generic;
     using System.IO;
-    using Tokens;
 
     /// <inheritdoc />
     /// <summary>
@@ -17,12 +16,7 @@
         public const string Symbol = "EI";
         
         /// <summary>
-        /// The tokens declared in order for this inline image object.
-        /// </summary>
-        public IReadOnlyList<IToken> ImageTokens { get; }
-
-        /// <summary>
-        /// The raw data for the inline image which should be interpreted according to the <see cref="ImageTokens"/>.
+        /// The raw data for the inline image which should be interpreted according to the corresponding <see cref="BeginInlineImageData.Dictionary"/>.
         /// </summary>
         public IReadOnlyList<byte> ImageData { get; }
         
@@ -32,17 +26,16 @@
         /// <summary>
         /// Create a new <see cref="EndInlineImage"/> operation.
         /// </summary>
-        /// <param name="imageTokens">The tokens which were set during the declaration of this image.</param>
         /// <param name="imageData">The raw byte data of this image.</param>
-        public EndInlineImage(IReadOnlyList<IToken> imageTokens, IReadOnlyList<byte> imageData)
+        public EndInlineImage(IReadOnlyList<byte> imageData)
         {
-            ImageTokens = imageTokens ?? throw new ArgumentNullException(nameof(imageTokens));
             ImageData = imageData ?? throw new ArgumentNullException(nameof(imageData));
         }
 
         /// <inheritdoc />
         public void Run(IOperationContext operationContext)
         {
+            operationContext.EndInlineImage(ImageData);
         }
 
         /// <inheritdoc />

src/UglyToad.PdfPig/Graphics/XObjectContentRecord.cs

@@ -1,6 +1,7 @@
 namespace UglyToad.PdfPig.Graphics
 {
     using System;
+    using Core;
     using PdfPig.Core;
     using Tokens;
     using Util.JetBrains.Annotations;
@@ -15,11 +16,15 @@
 
         public TransformationMatrix AppliedTransformation { get; }
 
-        public XObjectContentRecord(XObjectType type, StreamToken stream, TransformationMatrix appliedTransformation)
+        public RenderingIntent DefaultRenderingIntent { get; }
+
+        public XObjectContentRecord(XObjectType type, StreamToken stream, TransformationMatrix appliedTransformation,
+            RenderingIntent defaultRenderingIntent)
         {
             Type = type;
             Stream = stream ?? throw new ArgumentNullException(nameof(stream));
             AppliedTransformation = appliedTransformation;
+            DefaultRenderingIntent = defaultRenderingIntent;
         }
     }
 }

src/UglyToad.PdfPig/Parser/PageContentParser.cs

@@ -30,8 +30,23 @@
 
                 if (token is InlineImageDataToken inlineImageData)
                 {
-                    graphicsStateOperations.Add(BeginInlineImageData.Value);
-                    graphicsStateOperations.Add(new EndInlineImage(precedingTokens, inlineImageData.Data));
+                    var dictionary = new Dictionary<NameToken, IToken>();
+
+                    for (var i = 0; i < precedingTokens.Count - 1; i++)
+                    {
+                        var t = precedingTokens[i];
+                        if (!(t is NameToken n))
+                        {
+                            continue;
+                        }
+
+                        i++;
+
+                        dictionary[n] = precedingTokens[i];
+                    }
+
+                    graphicsStateOperations.Add(new BeginInlineImageData(dictionary));
+                    graphicsStateOperations.Add(new EndInlineImage(inlineImageData.Data));
                     precedingTokens.Clear();
                 }
                 else if (token is OperatorToken op)

src/UglyToad.PdfPig/Parser/PageFactory.cs

@@ -14,7 +14,6 @@
     using Tokenization.Scanner;
     using Tokens;
     using Util;
-    using XObjects;
 
     internal class PageFactory : IPageFactory
     {
@@ -22,18 +21,15 @@
         private readonly IResourceStore resourceStore;
         private readonly IFilterProvider filterProvider;
         private readonly IPageContentParser pageContentParser;
-        private readonly XObjectFactory xObjectFactory;
         private readonly ILog log;
 
         public PageFactory(IPdfTokenScanner pdfScanner, IResourceStore resourceStore, IFilterProvider filterProvider,
             IPageContentParser pageContentParser,
-            XObjectFactory xObjectFactory,
             ILog log)
         {
             this.resourceStore = resourceStore;
             this.filterProvider = filterProvider;
             this.pageContentParser = pageContentParser;
-            this.xObjectFactory = xObjectFactory;
             this.log = log;
             this.pdfScanner = pdfScanner;
         }
@@ -125,7 +121,7 @@
         {
             var operations = pageContentParser.Parse(new ByteArrayInputBytes(contentBytes));
 
-            var context = new ContentStreamProcessor(cropBox.Bounds, resourceStore, userSpaceUnit, rotation, isLenientParsing, pdfScanner, xObjectFactory, log);
+            var context = new ContentStreamProcessor(cropBox.Bounds, resourceStore, userSpaceUnit, rotation, isLenientParsing, pdfScanner, filterProvider, log);
 
             return context.Process(operations);
         }

src/UglyToad.PdfPig/Parser/PdfDocumentFactory.cs

@@ -26,7 +26,6 @@
     using Tokenization.Scanner;
     using Tokens;
     using Util;
-    using XObjects;
 
     internal static class PdfDocumentFactory
     {
@@ -123,11 +122,11 @@
                     new Type1FontParser(new Type1EncryptedPortionParser()), compactFontFormatParser),
                 new Type3FontHandler(pdfScanner, cMapCache, filterProvider, encodingReader));
             
-            var resourceContainer = new ResourceContainer(pdfScanner, fontFactory);
+            var resourceContainer = new ResourceStore(pdfScanner, fontFactory);
             
             var pageFactory = new PageFactory(pdfScanner, resourceContainer, filterProvider, 
                 new PageContentParser(new ReflectionGraphicsStateOperationFactory()), 
-                new XObjectFactory(), log);
+                log);
             var informationFactory = new DocumentInformationFactory();
 
             var information = informationFactory.Create(pdfScanner, crossReferenceTable.Trailer);

src/UglyToad.PdfPig/PdfDocument.cs

@@ -220,6 +220,7 @@
         {
             try
             {
+                pdfScanner.Dispose();
                 inputBytes.Dispose();
             }
             catch (Exception ex)

src/UglyToad.PdfPig/Tokenization/Scanner/IPdfTokenScanner.cs

@@ -0,0 +1,18 @@
+namespace UglyToad.PdfPig.Tokenization.Scanner
+{
+    using System;
+    using Tokens;
+
+    /// <summary>
+    /// Tokenizes objects from bytes in a PDF file.
+    /// </summary>
+    internal interface IPdfTokenScanner : ISeekableTokenScanner, IDisposable
+    {
+        /// <summary>
+        /// Tokenize the object with a given object number.
+        /// </summary>
+        /// <param name="reference">The object number for the object to tokenize.</param>
+        /// <returns>The tokenized object.</returns>
+        ObjectToken Get(IndirectReference reference);
+    }
+}

src/UglyToad.PdfPig/Tokenization/Scanner/PdfTokenScanner.cs

@@ -13,19 +13,6 @@
     using Parser.Parts;
     using Tokens;
 
-    /// <summary>
-    /// Tokenizes objects from bytes in a PDF file.
-    /// </summary>
-    internal interface IPdfTokenScanner : ISeekableTokenScanner
-    {
-        /// <summary>
-        /// Tokenize the object with a given object number.
-        /// </summary>
-        /// <param name="reference">The object number for the object to tokenize.</param>
-        /// <returns>The tokenized object.</returns>
-        ObjectToken Get(IndirectReference reference);
-    }
-
     internal class PdfTokenScanner : IPdfTokenScanner
     {
         private static readonly byte[] EndstreamBytes =
@@ -41,6 +28,7 @@
         private readonly CoreTokenScanner coreTokenScanner;
 
         private IEncryptionHandler encryptionHandler;
+        private bool isDisposed;
 
         /// <summary>
         /// Stores tokens encountered between obj - endobj markers for each <see cref="MoveNext"/> call.
@@ -75,6 +63,11 @@
 
         public bool MoveNext()
         {
+            if (isDisposed)
+            {
+                throw new ObjectDisposedException(nameof(PdfTokenScanner));
+            }
+
             // Read until we find object-number generation obj, e.g. "69 420 obj".
             int tokensRead = 0;
             while (coreTokenScanner.MoveNext() && !Equals(coreTokenScanner.CurrentToken, OperatorToken.StartObject))
@@ -576,26 +569,51 @@
 
         public bool TryReadToken<T>(out T token) where T : class, IToken
         {
+            if (isDisposed)
+            {
+                throw new ObjectDisposedException(nameof(PdfTokenScanner));
+            }
+
             return coreTokenScanner.TryReadToken(out token);
         }
 
         public void Seek(long position)
         {
+            if (isDisposed)
+            {
+                throw new ObjectDisposedException(nameof(PdfTokenScanner));
+            }
+
             coreTokenScanner.Seek(position);
         }
 
         public void RegisterCustomTokenizer(byte firstByte, ITokenizer tokenizer)
         {
+            if (isDisposed)
+            {
+                throw new ObjectDisposedException(nameof(PdfTokenScanner));
+            }
+
             coreTokenScanner.RegisterCustomTokenizer(firstByte, tokenizer);
         }
 
         public void DeregisterCustomTokenizer(ITokenizer tokenizer)
         {
+            if (isDisposed)
+            {
+                throw new ObjectDisposedException(nameof(PdfTokenScanner));
+            }
+
             coreTokenScanner.DeregisterCustomTokenizer(tokenizer);
         }
 
         public ObjectToken Get(IndirectReference reference)
         {
+            if (isDisposed)
+            {
+                throw new ObjectDisposedException(nameof(PdfTokenScanner));
+            }
+
             if (objectLocationProvider.TryGetCached(reference, out var objectToken))
             {
                 return objectToken;
@@ -717,5 +735,11 @@
 
             return results;
         }
+
+        public void Dispose()
+        {
+            inputBytes?.Dispose();
+            isDisposed = true;
+        }
     }
 }

src/UglyToad.PdfPig/Tokens/NameToken.Constants.cs

@@ -114,7 +114,7 @@
         public static readonly NameToken ColorDodge = new NameToken("ColorDodge");
         public static readonly NameToken Colorants = new NameToken("Colorants");
         public static readonly NameToken Colors = new NameToken("Colors");
-        public static readonly NameToken Colorspace = new NameToken("ColorSpace");
+        public static readonly NameToken ColorSpace = new NameToken("ColorSpace");
         public static readonly NameToken Columns = new NameToken("Columns");
         public static readonly NameToken Compatible = new NameToken("Compatible");
         public static readonly NameToken Components = new NameToken("Components");
@@ -272,6 +272,7 @@
         public static readonly NameToken Info = new NameToken("Info");
         public static readonly NameToken Ink = new NameToken("Ink");
         public static readonly NameToken Inklist = new NameToken("InkList");
+        public static readonly NameToken Intent = new NameToken("Intent");
         public static readonly NameToken Interpolate = new NameToken("Interpolate");
         public static readonly NameToken It = new NameToken("IT");
         public static readonly NameToken ItalicAngle = new NameToken("ItalicAngle");

src/UglyToad.PdfPig/UglyToad.PdfPig.csproj

@@ -23,6 +23,11 @@
     <AllowedOutputExtensionsInPackageBuildOutputFolder>$(AllowedOutputExtensionsInPackageBuildOutputFolder);.pdb</AllowedOutputExtensionsInPackageBuildOutputFolder>
   </PropertyGroup>
 
+  <PropertyGroup Condition="'$(Configuration)|$(TargetFramework)|$(Platform)'=='Debug|netstandard2.0|AnyCPU'">
+    <TreatWarningsAsErrors>true</TreatWarningsAsErrors>
+    <WarningsAsErrors />
+  </PropertyGroup>
+
   <ItemGroup>
     <None Remove="Resources\AdobeFontMetrics\*" />
     <None Remove="Resources\CMap\*" />

src/UglyToad.PdfPig/Util/OtherEncodings.cs

@@ -1,5 +1,7 @@
 namespace UglyToad.PdfPig.Util
 {
+    using System.Collections.Generic;
+    using System.Linq;
     using System.Text;
 
     internal static class OtherEncodings
@@ -19,6 +21,16 @@
             return Iso88591.GetBytes(s);
         }
 
+        public static string BytesAsLatin1String(IReadOnlyList<byte> bytes)
+        {
+            if (bytes is byte[] arr)
+            {
+                return BytesAsLatin1String(arr);
+            }
+
+            return BytesAsLatin1String(bytes.ToArray());
+        }
+
         public static string BytesAsLatin1String(byte[] bytes)
         {
             if (bytes == null)

src/UglyToad.PdfPig/XObjects/XObjectFactory.cs

@@ -1,13 +1,25 @@
 namespace UglyToad.PdfPig.XObjects
 {
     using System;
+    using System.Collections.Generic;
+    using System.Linq;
+    using Content;
+    using Exceptions;
+    using Filters;
+    using Geometry;
     using Graphics;
+    using Graphics.Colors;
+    using Graphics.Core;
     using Tokenization.Scanner;
     using Tokens;
+    using Util;
 
-    internal class XObjectFactory
+    internal static class XObjectFactory
     {
-        public XObjectImage CreateImage(XObjectContentRecord xObject, IPdfTokenScanner pdfScanner, bool isLenientParsing)
+        public static XObjectImage ReadImage(XObjectContentRecord xObject, IPdfTokenScanner pdfScanner,
+            IFilterProvider filterProvider,
+            IResourceStore resourceStore,
+            bool isLenientParsing)
         {
             if (xObject == null)
             {
@@ -19,19 +31,103 @@
                 throw new InvalidOperationException($"Cannot create an image from an XObject with type: {xObject.Type}.");
             }
 
-            var width = xObject.Stream.StreamDictionary.Get<NumericToken>(NameToken.Width, pdfScanner).Int;
-            var height = xObject.Stream.StreamDictionary.Get<NumericToken>(NameToken.Height, pdfScanner).Int;
+            var dictionary = xObject.Stream.StreamDictionary;
 
-            var isJpxDecode = xObject.Stream.StreamDictionary.TryGet(NameToken.Filter, out var token) 
+            var bounds = xObject.AppliedTransformation.Transform(new PdfRectangle(new PdfPoint(0, 0), new PdfPoint(1, 1)));
+
+            var width = dictionary.Get<NumericToken>(NameToken.Width, pdfScanner).Int;
+            var height = dictionary.Get<NumericToken>(NameToken.Height, pdfScanner).Int;
+
+            var isImageMask = dictionary.TryGet(NameToken.ImageMask, pdfScanner, out BooleanToken isMaskToken)
+                         && isMaskToken.Data;
+
+            var isJpxDecode = dictionary.TryGet(NameToken.Filter, out var token)
                 && token is NameToken filterName
                 && filterName.Equals(NameToken.JpxDecode);
 
-            var isImageMask = xObject.Stream.StreamDictionary.TryGet(NameToken.ImageMask, out var maskToken)
-                              && maskToken is BooleanToken maskBoolean
-                              && maskBoolean.Data;
+            int bitsPerComponent = 0;
+            if (!isImageMask && !isJpxDecode)
+            {
+                if (!dictionary.TryGet(NameToken.BitsPerComponent, pdfScanner, out NumericToken bitsPerComponentToken))
+                {
+                    throw new PdfDocumentFormatException($"No bits per component defined for image: {dictionary}.");
+                }
 
-            return new XObjectImage(width, height, isJpxDecode, isImageMask, xObject.Stream.StreamDictionary,
-                xObject.Stream.Data);
+                bitsPerComponent = bitsPerComponentToken.Int;
+            }
+            else if (isImageMask)
+            {
+                bitsPerComponent = 1;
+            }
+
+            var intent = xObject.DefaultRenderingIntent;
+            if (dictionary.TryGet(NameToken.Intent, out NameToken renderingIntentToken))
+            {
+                intent = renderingIntentToken.Data.ToRenderingIntent();
+            }
+
+            var interpolate = dictionary.TryGet(NameToken.Interpolate, pdfScanner, out BooleanToken interpolateToken)
+                              && interpolateToken.Data;
+
+            var decodedBytes = new Lazy<IReadOnlyList<byte>>(() => xObject.Stream.Decode(filterProvider));
+
+            var decode = EmptyArray<decimal>.Instance;
+
+            if (dictionary.TryGet(NameToken.Decode, pdfScanner, out ArrayToken decodeArrayToken))
+            {
+                decode = decodeArrayToken.Data.OfType<NumericToken>()
+                    .Select(x => x.Data)
+                    .ToArray();
+            }
+
+            var colorSpace = default(ColorSpace?);
+
+            if (!isImageMask)
+            {
+                if (dictionary.TryGet(NameToken.ColorSpace, pdfScanner, out NameToken colorSpaceNameToken)
+                    && TryMapColorSpace(colorSpaceNameToken, resourceStore, out var colorSpaceResult))
+                {
+                    colorSpace = colorSpaceResult;
+                }
+                else if (dictionary.TryGet(NameToken.ColorSpace, pdfScanner, out ArrayToken colorSpaceArrayToken))
+                {
+                    if (colorSpaceArrayToken.Length == 0)
+                    {
+                        throw new PdfDocumentFormatException($"Empty ColorSpace array defined for image XObject: {dictionary}.");
+                    }
+
+                    var first = colorSpaceArrayToken.Data[0];
+
+                    if (!(first is NameToken firstColorSpaceName) || !TryMapColorSpace(firstColorSpaceName, resourceStore, out colorSpaceResult))
+                    {
+                        throw new PdfDocumentFormatException($"Invalid ColorSpace array defined for image XObject: {colorSpaceArrayToken}.");
+                    }
+
+                    colorSpace = colorSpaceResult;
+                }
+                else if (!isJpxDecode)
+                {
+                    throw new PdfDocumentFormatException($"No ColorSpace defined for image XObject: {dictionary}.");
+                }
+            }
+
+            return new XObjectImage(bounds, width, height, bitsPerComponent, colorSpace, isJpxDecode, isImageMask, intent, interpolate, decode,
+                dictionary, xObject.Stream.Data, decodedBytes);
+        }
+
+        private static bool TryMapColorSpace(NameToken name, IResourceStore resourceStore, out ColorSpace colorSpaceResult)
+        {
+            if (name.TryMapToColorSpace(out colorSpaceResult))
+            {
+                return true;
+            }
+
+            if (!resourceStore.TryGetNamedColorSpace(name, out var colorSpaceNamedToken) || !(colorSpaceNamedToken is NameToken newName))
+            {
+                return false;
+            }
+
+            return newName.TryMapToColorSpace(out colorSpaceResult);
         }
     }
 }

src/UglyToad.PdfPig/XObjects/XObjectImage.cs

@@ -2,23 +2,35 @@
 {
     using System;
     using System.Collections.Generic;
+    using Content;
+    using Geometry;
+    using Graphics.Colors;
+    using Graphics.Core;
     using Tokens;
     using Util.JetBrains.Annotations;
 
+    /// <inheritdoc />
     /// <summary>
-    /// The raw stream from a PDF document representing an image XObject.
+    /// A PostScript image XObject.
     /// </summary>
-    public class XObjectImage
+    public class XObjectImage : IPdfImage
     {
-        /// <summary>
-        /// The width of the image in samples.
-        /// </summary>
-        public int Width { get; }
+        private readonly Lazy<IReadOnlyList<byte>> bytes;
 
-        /// <summary>
-        /// The height of the image in samples.
-        /// </summary>
-        public int Height { get; }
+        /// <inheritdoc />
+        public PdfRectangle Bounds { get; }
+
+        /// <inheritdoc />
+        public int WidthInSamples { get; }
+
+        /// <inheritdoc />
+        public int HeightInSamples { get; }
+
+        /// <inheritdoc />
+        public ColorSpace? ColorSpace { get; }
+
+        /// <inheritdoc />
+        public int BitsPerComponent { get; }
 
         /// <summary>
         /// The JPX filter encodes data using the JPEG2000 compression method.
@@ -27,41 +39,67 @@
         /// </summary>
         public bool IsJpxEncoded { get; }
 
-        /// <summary>
-        /// Whether this image should be treated as an image maske.
-        /// </summary>
+        /// <inheritdoc />
+        public RenderingIntent RenderingIntent { get; }
+
+        /// <inheritdoc />
         public bool IsImageMask { get; }
 
+        /// <inheritdoc />
+        public IReadOnlyList<decimal> Decode { get; }
+
+        /// <inheritdoc />
+        public bool Interpolate { get; }
+
+        /// <inheritdoc />
+        public bool IsInlineImage { get; } = false;
+
         /// <summary>
         /// The full dictionary for this Image XObject.
         /// </summary>
         [NotNull]
         public DictionaryToken ImageDictionary { get; }
 
-        /// <summary>
-        /// The encoded bytes of this image, must be decoded via any
-        /// filters defined in the <see cref="ImageDictionary"/> prior to consumption.
-        /// </summary>
+        /// <inheritdoc />
+        public IReadOnlyList<byte> RawBytes { get; }
+
+        /// <inheritdoc />
         [NotNull]
-        public IReadOnlyList<byte> Bytes { get; }
+        public IReadOnlyList<byte> Bytes => bytes.Value;
         
         /// <summary>
         /// Creates a new <see cref="XObjectImage"/>.
         /// </summary>
-        internal XObjectImage(int width, int height, bool isJpxEncoded, bool isImageMask, DictionaryToken imageDictionary, IReadOnlyList<byte> bytes)
+        internal XObjectImage(PdfRectangle bounds, int widthInSamples, int heightInSamples, int bitsPerComponent,
+            ColorSpace? colorSpace,
+            bool isJpxEncoded,
+            bool isImageMask,
+            RenderingIntent renderingIntent,
+            bool interpolate,
+            IReadOnlyList<decimal> decode,
+            DictionaryToken imageDictionary,
+            IReadOnlyList<byte> rawBytes, 
+            Lazy<IReadOnlyList<byte>> bytes)
         {
-            Width = width;
-            Height = height;
+            Bounds = bounds;
+            WidthInSamples = widthInSamples;
+            HeightInSamples = heightInSamples;
+            BitsPerComponent = bitsPerComponent;
+            ColorSpace = colorSpace;
             IsJpxEncoded = isJpxEncoded;
             IsImageMask = isImageMask;
+            RenderingIntent = renderingIntent;
+            Interpolate = interpolate;
+            Decode = decode;
             ImageDictionary = imageDictionary ?? throw new ArgumentNullException(nameof(imageDictionary));
-            Bytes = bytes ?? throw new ArgumentNullException(nameof(bytes));
+            RawBytes = rawBytes;
+            this.bytes = bytes ?? throw new ArgumentNullException(nameof(bytes));
         }
 
         /// <inheritdoc />
         public override string ToString()
         {
-            return ImageDictionary.ToString();
+            return $"XObject Image (w {Bounds.Width}, h {Bounds.Height}): {ImageDictionary}";
         }
     }
 }