YaoFANGUK/video-subtitle-extractor
Video Subtitle Extractor
This project is an optical character recognition tool designed to extract hardcoded subtitles from video frames and convert them into synchronized subtitle files. It functions as a text processor that transforms embedded visual text into a written format to improve video accessibility and translation.
The system uses graphics processing units to increase the speed and accuracy of text recognition. It includes a subtitle cleaning tool that applies custom mapping configurations to filter out watermarks, channel logos, and duplicate lines from the extracted text.
The tool supports batch processing for multiple video files that share identical resolutions and text region settings. It utilizes region-based extraction to isolate subtitles from background noise and synchronizes recognized text strings with specific video timestamps.
Features
- Optical Character Recognition - Uses optical character recognition to convert visual text found within video frames into digital strings.
- Hardcoded Subtitle Extractors - Extracts embedded text from video frames and saves it as synchronized written subtitle files.
- GPU-Accelerated Inference - Offloads text recognition workloads to graphics hardware to reduce processing time and increase throughput.
- GPU Acceleration - Uses GPU acceleration to increase the speed and accuracy of the text recognition process.
- OCR Acceleration - Utilizes graphics processing units to significantly increase the throughput and accuracy of subtitle text recognition.
- GPU-Accelerated Text Processing - Implements GPU-accelerated processing to reduce total text recognition time.
- Automated Video Subtitling - Automates the pipeline of extracting and aligning subtitles from multiple video files.
- Video Transcript Extraction - Extracts subtitles and timing data from video frames to create written transcripts for accessibility.
- Subtitle Processing - Implements a full pipeline for extracting, synchronizing, and processing hardcoded subtitles from video content.
- Hardcoded Subtitle Extraction - Converts text found within video frames into synchronized files using local optical character recognition.
- Hardcoded Subtitle Extraction - Converts text embedded directly into video frames into editable subtitle files using local optical character recognition.
- Coordinate-Based Extraction - Isolates subtitles from background noise by focusing analysis on specific coordinate areas within video frames.
- Temporal Subtitle Synchronization - Maps recognized text strings to specific video timestamps to generate synchronized subtitle files.
- Automated Subtitle Cleaning - Removes watermarks, logos, and recurring noise from extracted text using custom mapping configurations.
- Video Accessibility Tools - Creates written transcripts and synchronized subtitles for videos to improve viewer accessibility.
- Text Cleaning Pipelines - Standardizes extracted text by removing duplicates and visual noise through a cleaning pipeline.
- Subtitle Noise Reduction - Removes watermarks, channel logos, and duplicate lines from extracted text using custom mapping configurations.
- Watermark and Logo Filtering - Filters out watermarks, channel logos, and duplicate lines using a custom mapping configuration.
- Batch Video Processing - Enables subtitles to be extracted from multiple video files simultaneously when resolution and text regions are identical.
- Video File Processors - Provides capabilities to process multiple video files for text extraction based on shared resolution and region settings.
- Watermark Removal - Removes visual watermarks and channel logos from the final text output.
