Aspose.OCR for .NET 23.3.1 - Release Notes

Deprecation warning

What was changed

KeySummaryCategory
OCRNET‑636
OCRNET‑646
A slimmer, faster and more straightforward API has been introduced. See Added public APIs for details.New feature
OCRNET‑648Most of the existing API classes and methods have been marked as deprecated to remind you to update your existing code. They remain functional but will be removed in release 23.11.0 (November 2023) in favor of the new API introduced in this release. See Deprecated APIs for details.Enhancement

Public API changes and backwards compatibility

This section lists all public API changes introduced in Aspose.OCR for .NET 23.3.1 that may affect the code of existing applications.

Added public APIs:

The following public APIs have been introduced in this release:

Aspose.OCR.OcrInput class

The universal class for providing any type of data (images, PDF documents, archives, folders, streams, arrays, and so on) to the new image processing and recognition methods.

Aspose.OCR.ImageProcessing class

Specially adjust one or more files to improve the accuracy and reliability of the OCR. This class provides extended replacements for Aspose.OCR.AsposeOcr.PreprocessImage method:

MethodAction
Save(OcrInput images, string folderPath)Saves processed images to a folder. Replaces PreprocessImage method.
Render(OcrInput images)Processes files and returns an OcrInput object with adjusted images that can be passed to recognition methods.

Aspose.OCR.AsposeOcr.Recognize method

Recognize one or more files provided as an OcrInput object. It is a universal replacement for the following recognition methods:

MethodAction
RecognizeImageExtract text from a raster image, provided as file, memory stream, or a pixel array.
RecognizePdfExtract text from a PDF document.
RecognizeTiffExtract text from a multi-page TIFF image.
RecognizeDjvuExtract text from a DjVu file.
RecognizeImageFromUriRecognize an image hosted on website without downloading it to the computer.
RecognizeMultipleImagesBatch recognition.
RecognizeImageFromBase64Extract text from Base64 encoded images.

Aspose.OCR.AsposeOcr.RecognizeLines method

Recognize files containing a single line of text in the fastest possible mode. It is an extended replacement for RecognizeLine method.

Aspose.OCR.AsposeOcr.DetectRectangles method

Find areas of images containing text. It is an extended replacement for GetRectangles method.

Aspose.OCR.AsposeOcr.CalculateSkew method

Find out skew angles of provided images. It is an universal replacement for the following methods:

MethodAction
CalculateSkewDetect the skew angle of an image.
CalculateSkewFromUriDetect the skew angle of an image hosted on website without downloading it to the computer.

AllowedSymbols recognition setting

Limit a subset of recognized characters instead of using all symbols from the selected language. It is a replacement of the alphabet argument of AsposeOcr constructor and AllowedCharacters recognition setting.

IgnoredSymbols recognition setting

A blacklist of characters that are ignored during recognition. It is a replacement of the IgnoredCharacters recognition setting to ensure the consistent naming.

Updated public APIs:

No changes.

Removed public APIs:

The following public APIs have been removed in this release:

detectAreas argument of RecognitionSettings constructor

RecognitionSettings constructor no longer accepts detectAreas argument. Specify the area detection mode in DetectAreasMode recognition setting instead.

Deprecated APIs

The following public APIs have been marked as deprecated and will be removed in 23.11.0 (November 2023) release:

CalculateSkew method

Replaced with Aspose.OCR.AsposeOcr.CalculateSkew method.

CalculateSkewFromUri method

Replaced with Aspose.OCR.AsposeOcr.CalculateSkew method.

Aspose.OCR.AsposeOcr.PreprocessImage method

Replaced with Save method of Aspose.OCR.AsposeOcr.CalculateSkew class.

RecognizeImage

Replaced with Aspose.OCR.AsposeOcr.Recognize method.

RecognizePdf

Replaced with Aspose.OCR.AsposeOcr.Recognize method.

RecognizeTiff

Replaced with Aspose.OCR.AsposeOcr.Recognize method.

RecognizeDjvu

Replaced with Aspose.OCR.AsposeOcr.Recognize method.

RecognizeImageFromUri

Replaced with Aspose.OCR.AsposeOcr.Recognize method.

RecognizeMultipleImages

Replaced with Aspose.OCR.AsposeOcr.Recognize method.

RecognizeImageFromBase64

Replaced with Aspose.OCR.AsposeOcr.Recognize method.

RecognizeLine

Replaced with Aspose.OCR.AsposeOcr.RecognizeLines method.

GetRectangles

Replaced with Aspose.OCR.AsposeOcr.DetectRectangles method.

CalculateSkew

Replaced with Aspose.OCR.AsposeOcr.CalculateSkew(OcrInput images) override.

CalculateSkewFromUri

Replaced with Aspose.OCR.AsposeOcr.CalculateSkew(OcrInput images) method.

DocumentRecognitionSettings

No longer required. Use RecognitionSettings class as a universal replacement.

AutoSkew recognition setting

No longer required. Use AutoSkew image processing filter instead.

SkewAngle recognition setting

No longer required. Use Rotate image processing filter instead.

ThresholdValue recognition setting

No longer required. Use the binarization threshold setting in Binarize image processing filter instead.

AutoContrast recognition setting

No longer required. Use ContrastCorrectionFilter image processing filter instead.

AutoDenoising recognition setting

No longer required. Use AutoDenoising image processing filter instead.

PreprocessingFilters recognition setting

No longer used in recognition methods. Process images before proceeding with recognition or provide processing filters in OcrInput object.

AllowedCharacters recognition setting

No longer required. Use the new AllowedSymbols recognition setting instead.

IgnoredCharacters recognition setting

No longer required. Use the new IgnoredSymbols recognition setting instead.

AsposeOcr(string alphabet) constructor

No longer required. Define the list of allowed characters through the new AllowedSymbols recognition setting instead.

Examples

The examples below illustrates the changes introduced in this release:

Migrating to the new API

Original code (Aspose.OCR for .NET 23.2.1 and below):

Aspose.OCR.AsposeOcr recognitionEngine = new Aspose.OCR.AsposeOcr();
// Correct geometric distortions
Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter filters = new Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter();
filters.Add(Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter.AutoDewarping());
// Convert the first 3 pages of PDF to searchable PDF
Aspose.OCR.DocumentRecognitionSettings recognitionSettings1 = new Aspose.OCR.DocumentRecognitionSettings();
recognitionSettings1.Language = Aspose.OCR.Language.Ukr;
recognitionSettings1.AutoContrast = true;
recognitionSettings1.StartPage = 0;
recognitionSettings1.PagesNumber = 3;
List<Aspose.OCR.RecognitionResult> results = recognitionEngine.RecognizePdf("source1.pdf", recognitionSettings1);
Aspose.OCR.AsposeOcr.SaveMultipageDocument("result1.pdf", Aspose.OCR.SaveFormat.Pdf, results);
// Convert the second PDF to searchable PDF
Aspose.OCR.DocumentRecognitionSettings recognitionSettings2 = new Aspose.OCR.DocumentRecognitionSettings();
recognitionSettings2.Language = Aspose.OCR.Language.Ukr;
recognitionSettings2.AutoContrast = true;
List<Aspose.OCR.RecognitionResult> results = recognitionEngine.RecognizePdf("source2.pdf", recognitionSettings2);
Aspose.OCR.AsposeOcr.SaveMultipageDocument("result2.pdf", Aspose.OCR.SaveFormat.Pdf, results);

New code (Aspose.OCR for .NET 23.3.1 and above):

Aspose.OCR.AsposeOcr recognitionEngine = new Aspose.OCR.AsposeOcr();
Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter filters = new Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter();
filters.Add(Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter.AutoDewarping());
// Activate automatic contrast adjustment in processing filters instead of recognition settings
filters.Add(Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter.ContrastCorrectionFilter());
// Add all PDF documents to OcrInput object and apply processing filters
Aspose.OCR.OcrInput input = new Aspose.OCR.OcrInput(InputType.PDF, filters);
// Specify page numbers when adding a file to OcrInput
input.Add("source1.pdf", 0, 3);
input.Add("source2.pdf");
// Use RecognitionSettings instead of DocumentRecognitionSettings
Aspose.OCR.RecognitionSettings recognitionSettings = new Aspose.OCR.RecognitionSettings();
// Remove automatic contrast adjustment from recognition settings - it is already applied through image processing filters
recognitionSettings.Language = Aspose.OCR.Language.Ukr;
// Recognize all files in one universal call
List<Aspose.OCR.RecognitionResult> results = recognitionEngine.Recognize(input, recognitionSettings);
// Save recognition results as searchable PDFs
for(int i=0;i<results.count;i++) Aspose.OCR.AsposeOcr.SaveMultipageDocument($"result{i+1}.pdf", Aspose.OCR.SaveFormat.Pdf, results[i]);

Process and save all images from PDF documents

// Set processing filters
Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter filters = new Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter();
filters.Add(Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter.AutoDewarping());
filters.Add(Aspose.OCR.Models.PreprocessingFilters.PreprocessingFilter.ContrastCorrectionFilter());
// Add all PDF documents to OcrInput object and apply processing filters
Aspose.OCR.OcrInput input = new Aspose.OCR.OcrInput(InputType.PDF, filters);
input.Add("source1.pdf", 0, 3);
input.Add("source2.pdf");
// Save all images from provided PDFs to the folder
Aspose.OCR.ImageProcessing.Save(input, @"C:\images");

Detect skew angles

Aspose.OCR.AsposeOcr recognitionEngine = new Aspose.OCR.AsposeOcr();
// Add all PDF documents to OcrInput object
Aspose.OCR.OcrInput input = new Aspose.OCR.OcrInput(InputType.PDF);
input.Add("source1.pdf", 0, 3);
input.Add("source2.pdf");
// Detect skew angles
List<Aspose.OCR.SkewOutput> angles = recognitionEngine.CalculateSkew(input);
foreach(Aspose.OCR.SkewOutput angle in angles) Console.WriteLine($"{angle.Source} {angle.Page} {angle.ImageIndex} {angle.Angle}");