private IPix Preprocess(IPix pix)
{
IPix result = null;
foreach (var preprocessor in _preprocessors)
{
result = preprocessor.Run(pix);
}
return(result);
}
public IPix PrepareOneBitPerPixel(IPix pix)
{
if (pix == null)
{
throw new ArgumentNullException(nameof(pix));
}
var pointer = Leptonica5Filters.pixPrepare1bpp(pix.HandleRef.Handle, IntPtr.Zero, 0, 0).GetPointerOrThrow();
return(new Pix(pointer));
}
public IPix OrientationCorrect(IPix pix, float minUpConfidence = 4f, float minRatio = 2.5f)
{
if (pix == null)
{
throw new ArgumentNullException(nameof(pix));
}
var pointer = Leptonica5Filters.pixOrientCorrect(pix.HandleRef.Handle, minUpConfidence, minRatio, out _, out _, out _, 0).GetPointerOrThrow();
return(new Pix(pointer));
}
public IBox FindPageForeground(IPix pix, int threshold = 128, int minDistance = 50, int eraseDistance = 70)
{
if (pix == null)
{
throw new ArgumentNullException(nameof(pix));
}
var pointer = Leptonica5Filters.pixFindPageForeground(pix.HandleRef.Handle, threshold, minDistance, eraseDistance, 0, IntPtr.Zero).GetPointerOrThrow();
return(new Box(pointer));
}
public IPix ClipRectangle(IPix pix, IBox box)
{
if (pix == null)
{
throw new ArgumentNullException(nameof(pix));
}
var pointer = Leptonica5Filters.pixClipRectangle(pix.HandleRef.Handle, box.HandleRef.Handle, IntPtr.Zero).GetPointerOrThrow();
return(new Pix(pointer));
}
public IPix DeskewBoth(IPix pix, DeskewReductionFactor searchReduction)
{
if (pix == null)
{
throw new ArgumentNullException(nameof(pix));
}
var pointer = Leptonica5Filters.pixDeskewBoth(pix.HandleRef.Handle, (int)searchReduction).GetPointerOrThrow();
return(new Pix(pointer));
}
public void AddPix(IPix pix)
{
if (pix == null)
{
throw new ArgumentNullException(nameof(pix));
}
if (Leptonica5Pix.pixaAddPix(HandleRef.Handle, pix.HandleRef.Handle, 1) != 0)
{
throw new InvalidOperationException("Leptonica failed.");
}
}
public override IPix Run(IPix image)
{
if (image == null)
{
throw new ArgumentNullException(nameof(image));
}
using var deskewedPix = leptonicaInterop.DeskewBoth(image, ReductionFactor);
using var oneBitPix = leptonicaInterop.PrepareOneBitPerPixel(deskewedPix);
using var correctedPix = leptonicaInterop.OrientationCorrect(oneBitPix, MinUpConfidence, MinRatio);
using var foregroundBox = leptonicaInterop.FindPageForeground(correctedPix, Threshold, MinDistance, EraseDistance);
return(leptonicaInterop.ClipRectangle(correctedPix, foregroundBox));
}
public override IPix Run(IPix image)
{
if (image == null)
{
throw new ArgumentNullException(nameof(image));
}
return(image
.ChainOutput(x => leptonicaInterop.DeskewBoth(x, ReductionFactor), disposePreviousPix: false)
.ChainOutput(x => leptonicaInterop.PrepareOneBitPerPixel(x))
.ChainOutput(x => leptonicaInterop.OrientationCorrect(x, MinUpConfidence, MinRatio))
.ChainOutput(x =>
{
using var box = leptonicaInterop.FindPageForeground(x, Threshold, MinDistance, EraseDistance);
return leptonicaInterop.ClipRectangle(x, box);
}));
}
public static IPix ChainOutput(this IPix pix, Func <IPix, IPix> invocation, bool disposePreviousPix = true)
{
if (pix == null)
{
throw new ArgumentNullException(nameof(pix));
}
if (invocation == null)
{
throw new ArgumentNullException(nameof(invocation));
}
if (!disposePreviousPix)
{
return(invocation(pix));
}
using (pix)
return(invocation(pix));
}
/// <summary>
/// Processes the specific image.
/// </summary>
/// <remarks>
/// You can only have one result iterator open at any one time.
/// </remarks>
/// <param name="image">The image to process.</param>
/// <param name="region">The image region to process.</param>
/// <returns>A result iterator</returns>
public ResultIterator Process(IPix image, Rect region)
{
if (image == null) throw new ArgumentNullException("image");
if (region.X1 < 0 || region.Y1 < 0 || region.X2 > image.Width || region.Y2 > image.Height)
throw new ArgumentException("The image region to be processed must be within the image bounds.", "region");
if (processCount > 0) throw new InvalidOperationException("Only one image can be processed at once. Please make sure you dispose of the result iterator once your finished with it.");
processCount++;
Interop.TessApi.BaseApiSetImage(handle, image.Handle);
Interop.TessApi.BaseApiSetRectangle(handle, region.X1, region.Y1, region.Width, region.Height);
if (Interop.TessApi.BaseApiRecognize(handle, IntPtr.Zero) != 0) {
throw new InvalidOperationException("failed to process document.");
}
var iteratorHandle = Interop.TessApi.BaseApiGetIterator(handle);
var iterator = new ResultIterator(iteratorHandle);
iterator.Disposed += OnIteratorDisposed;
return iterator;
}
public virtual Func <Tr> PixStack(PixStack ps, Func <Tr>[] subArray)
{
return(() =>
{
var typedOther = Other as PixStack;
if (typedOther == null)
{
throw new ArgumentException();
}
var length = subArray.Length;
if (length != typedOther.PixArray.Length)
{
throw new ArgumentException();
}
var result = ProductOp.PixStack(ps, typedOther,
subArray.ProductArray(typedOther.PixArray, length,
(fp0, p1) => { Other = p1; return fp0(); }));
Other = typedOther;
return result;
});
}
public ITesseractResult Process(IPix image)
{
if (image == null)
{
throw new ArgumentNullException(nameof(image));
}
if (Interlocked.CompareExchange(ref _processCount, 1, 0) != 0)
{
throw new InvalidOperationException("Only one image can be processed at once. Please make sure you dispose of the page once your finished with it.");
}
try
{
const PageSegmentMode pageSegMode = PageSegmentMode.Auto;
Tesseract4.TessBaseAPISetPageSegMode(_handle, (int)pageSegMode);
Tesseract4.TessBaseAPISetImage2(_handle, image.HandleRef);
if (Tesseract4.TessBaseAPIRecognize(_handle, out var rec) != 0)
{
throw new TesseractException("Recognition of image failed.");
}
using var text = Text.Create(() => Tesseract4.TessBaseAPIGetUTF8Text(_handle));
using var hocrText = Text.Create(() => Tesseract4.TessBaseAPIGetHOCRText(_handle, 0));
return(new TesseractResult
{
Text = text.ToString(),
HocrText = $"{Tags.XhtmlBeginTag}{hocrText}{Tags.XhtmlEndTag}",
Confidence = null
});
}
finally
{
Tesseract4.TessBaseAPIClear(_handle);
Interlocked.Decrement(ref _processCount);
}
}
public PixSaveOp(IPix other, string pathPrefix)
: base(new PixIntSumProductOp(), other)
{
PathPrefix = pathPrefix;
}
public static IPix WithRemovedMipMaps(this IPix pix)
{
return(pix.Op(new PixRemoveMipMaps()));
}
public static IPix WithLoadedImages(this IPix pix, string pathPrefix = null)
{
return(pix.Op(new PixLoadOp(pathPrefix == null ? "" : pathPrefix)));
}
public static IPix WithMipMaps(
this IPix pix, PixImageMipMap.MipMapOptions options = null)
{
return(pix.Op(new PixCreateMipMaps(options == null
? PixImageMipMap.MipMapOptions.Default : options)));
}
public ResultIterator Process(IPix image)
{
return Process(image, new Rect(0, 0, image.Width, image.Height));
}
public static PixImage[] GetPixImageArray(this IPix pix, int level = 0)
{
return(pix.Op(new PixGetPixImageArray(level))());
}
public static IPix GetDescriptors(this IPix pix)
{
return(pix.Op(new PixGetInfo()));
}
public static IPix ToPixImage <T>(this IPix pix)
{
return(pix.Op(new PixToPixImage <T>()));
}
public PixInnerProductOp(IPixProductOp <Tr> productOp, IPix other)
{
ProductOp = productOp;
Other = other;
}
public abstract IPix Run(IPix image);
/// <summary>
/// </summary>
/// <returns>The number of saved PixImages.</returns>
public static int SaveImages(this IPix pix, IPix names, string pathPrefix = null)
{
return(pix.Op(new PixSaveOp(names, pathPrefix == null ? "" : pathPrefix))());
}