Released: Nov 9, 2022

Updates in V22.10

Funktionen

• Extract text from an in-memory (InputStream) TIFF image.
• Extract text from an in-memory (BufferedImage) receipt image.
• PDF recognition method can now be called from AsposeOCR class.

Released: Sep 28, 2022

Updates in V22.9

Funktionen

• You can now extract text from receipts.
• New setting improves recognition of small fonts and detection of dense lines.
• Improved handling of images rotated by 90 degrees.
• Original font size is now preserved in OCR results saved in PDF and DOCX formats.

Released: Aug 29, 2022

Updates in V22.8

Funktionen

• Added support for recognizing text from images provided as an array of bytes.
• Preprocessing filters can now be applied to specific areas of an image.

Fehlerkorrekturen

• Aspose.OCR for Java is now displayed in the properties of generated PDF documents instead of Aspose.OCR for .NET.
• Fixed black page output when recognizing scanned PDF files.

Released: Jul 25, 2022

Updates in V22.7

Funktionen

• Added image noise reduction filter.
• Added image skew correction filter.

Released: Jul 1, 2022

Updates in V22.6

Funktionen

• Added a new detection mode for better identification and recognition of tabular structures.
• Added a new machine learning model for image denoising:
  • Integrated Binarized Neural Network (BNN) and related tests.
  • Added pre- and post-processing algorithms for Binarized Neural Network.

Fehlerkorrekturen

• Fixed an issue with saving recognition results in PDF format.

Released: May 20, 2022

Updates in V22.5

Funktionen

• Added support for the Cyrillic alphabet along with Belorussian, Bulgarian, Ukrainian, Kazakh, Russian and Serbian languages.
• Added the ability to save results in XML or JSON file format.
• New API method improves recognition speed.

Released: Apr 21, 2022

Updates in V22.4

Funktionen

• Improved algorithm for creating XLSX files.
• Improved algorithm for creating text files.
• Added the ability to load a PDF from an input stream.
• Updated ONNX runtime to version 1.11.

Released: Mar 31, 2022

Updates in V22.3

Funktionen

• Improved text area detection.
• Added the ability to get the coordinates of words.
• Extended recognition result object data.

Released: Feb 21, 2022

Updates in V22.2

Funktionen

• Added the ability to create a combine mode for DSR/CRAFT areas recognition.
• Integrated algorithms for CRAFT into Ocr.Java project.
• Extended the API - Added new enum DetectAreasMode with fields: DOCUMENT, PHOTO, COMBINED and NONE.

Fehlerkorrekturen

• Fixed bug for import optimized models.
• Fixed bug for postprocessing in the image recognition process with the Craft-model.

Released: Dec 24, 2021

Updates in V21.12

Funktionen

• Improved skew correction algorithm.
• Added the ability to create XLSX (Microsoft Excel) documents as a result of image recognition.
• Added the ability to create searchable PDF and DOCX documents with original text structure.