Aspose.OCR for Python via .NET 25.6.0 - Release Notes

What was changed

Public API changes and backwards compatibility

This section lists all public API changes introduced in Aspose.OCR for Python via .NET 25.6.0 that June affect the code of existing applications.

Added public APIs:

The following public APIs have been introduced in this release:

Debug mode

You can save intermediate image processing results — such as preprocessed images and text detection overlays — for visual inspection and troubleshooting. Debug mode is enabled through set methods of the AsposeOcr class:

Confidence score

Each recognized text line includes an optional confidence score — a floating-point value between 0.0 and 1.0 stored in the confidence field of the LinesResult class.

This score reflects the recognition certainty of the line:

• 1.0 — the engine is completely confident the recognition is correct.
• 0.0 — recognition confidence is unknown or not calculated.

⚠️ The value is always set to 0.0 when using a temporary license.

The confidence score is only calculated for specific languages:

✅ Supported: Chinese (all groups), Arabic, Hindi, European, Korean, Japanese, Telugu, Tamil, Kannada

❌ Not supported: ExtLatin or languages with diacritical marks

You can use this value to filter or highlight low-confidence results in your application.

Updated public APIs:

No changes.

Deprecated APIs

The following public APIs have been marked as deprecated and will be removed in 25.10.0 (October 2025) release:

RectangleOutput class

AsposeOcr.detect_rectangles method

RecognitionResult.recognition_areas_text

RecognitionResult.recognition_areas_rectangles

RecognitionResult.skew

CharacterRecognitionResult.image_index

SkewOutput.image_index

RecognitionResult.skew

RecognitionResult.skew

RecognitionResult.skew

Removed public APIs:

No changes.

Examples

The code samples below illustrate the changes introduced in this release:

Enable Debug Mode for Logging

from aspose.ocr.models import *


# Initialize recognition API
api = new AsposeOcr()

# Enable debug mode to log internal processing information
api.set_debug_mode_save_directory("D:\\output\\debug")
api.set_debug_mode(True)

# Add an image to OcrInput object
input = new OcrInput(InputType.SINGLE_IMAGE)
input.add("source.png")

# Recognize image
results = api.recognize(input)

Get Confidence of Recognized Text

from aspose.ocr.models import *

# Initialize recognition API
api = new AsposeOCR();

# Add an image to OcrInput object
input = new OcrInput(InputType.SINGLE_IMAGE)
input.add("source.png");

# Recognize image
OcrOutput results = api.recognize(input);

# Print recognized text with confidence
for result in results:
    print("Text:", result.recognition_text)

    for line in result.recognition_lines_result:
        print(f"{line.text_in_line} — Confidence: {line.confidence}")