Craft-Text Detector

CRAFT-Text Detector is an open-source Python library designed to detect text regions in images using the Character Region Awareness for Text (CRAFT) deep learning model. It accurately identifies text at the character level, making it highly effective for natural scene images, scanned documents, and OCR preprocessing tasks.

Features of CRAFT-Text Detector:

• Detects text regions with high precision at the character level.
• Supports images of various sizes and resolutions.
• Compatible with both CPU and GPU (CUDA) execution.
• Provides bounding box coordinates for detected text regions.
• Easy integration with OCR pipelines and machine learning workflows.
• Python-based and suitable for research and production environments.
• Open-source and customizable for various computer vision applications.

CRAFT-Text Detector Usage

$ sudo su
$ source /opt/craft-text-detector/venv/bin/activate
$ pip show craft-text-detector | grep Version