aygong / video-ocr

video-ocr

video-ocr is a command line tool and a python library that performs OCR on video frames, reducing the computational effort by choosing only frames that are different from their adjacent frames.

Installation

1. video-ocr uses tesserocr, which in turn uses tesseract as the OCR engine. So, please follow the tesserocr installation instructions.
2. Install using pip: pip install video-ocr

Getting started

Command line usage

Run video-ocr --help to show the help text. Example usage is

video-ocr /path/to/video.mp4 --sample_rate=1 --debug_dir="path/to/debug/dir"

Python usage

import video_ocr
video_ocr.perform_video_ocr("/Users/veneetreddy/Downloads/texts.mp4", sample_rate=1)