Skip to content

This tool is designed to identify potential issues in videos where rolling text at the bottom of the screen intersects with subtitles. It utilizes easyocr to detect text in frames and creates bounding boxes. The program checks for the presence of subtitles in each frame and calculates the intersection between the bounding box and subtitle area.

Notifications You must be signed in to change notification settings

Snimm/subtitle-quality-quantification-tool

Repository files navigation

Project Name

Subtitle Analysis Tool

Overview

This tool provides functionality for analysing subtitle. It checks if the subtitle and the text present in the video overlap. It includes two main commands: main for analyzing videos for subtitle obstruction and get-video-details for extracting video details such as FPS, frame count, width, and height.

Installation

Make sure you have Python installed on your system. It is recommended to use a virtual environment. Here are the installation steps:

# Clone the repository
git clone https://github.com/Snimm/subtitle-quality-quantification-tool

# Navigate to the project directory
cd subtitle-quality-quantification-tool

# Create and activate a virtual environment (optional but recommended)
python -m venv .venv
source .venv/bin/activate  # For Linux/macOS
# or
.\.venv\Scripts\activate  # For Windows

# Install the required dependencies
pip install -r requirements.txt

# Run the application
python main.py --help

Usage

main Command

The main command analyzes a video for subtitle obstruction.

python main.py main [OPTIONS] VIDEO_PATH SUB_PATH

Arguments:

  • VIDEO_PATH [required]: Path to the video file.
  • SUB_PATH [required]: Path to the subtitle file.

Options:

  • --percentage-width-covered-by-sub FLOAT: Percentage of video width covered by subtitles (default: 85).
  • --percentage-height-covered-by-sub FLOAT: Percentage of video height covered by subtitles (default: 15). It is assumed that the subtitles are present at the very bottom of the video.
  • --display-images-with-issue / --no-display-images-with-issue: Display images with subtitle obstruction (default: display-images-with-issue).
  • --save-images-with-issue / --no-save-images-with-issue: Save images with subtitle obstruction (default: no-save-images-with-issue). Images are saved in position_issue_frames folder.
  • --number-of-frames-to-analyze INTEGER: Number of frames to analyze (default: 10). Frames that are analysed are distributed evenly across the video.
  • --help: Show this message and exit.

get-video-details Command

The get-video-details command extracts video details such as FPS, frame count, width, and height.

python main.py get-video-details [OPTIONS] VIDEO_PATH

Arguments:

  • VIDEO_PATH [required]: Path to the video file.

Options:

  • --help: Show this message and exit.

Examples

Analyze a video for subtitle obstruction:

python main.py main --percentage-width-covered-by-sub 80 --percentage-height-covered-by-sub 10 --save-images-with-issue --number-of-frames-to-analyze 5 video.mp4 subtitles.srt

Extract video details:

python main.py get-video-details video.mp4

Support and Contribution

Feel free to open an issue for any problems or suggestions. Contributions are welcome!

About

This tool is designed to identify potential issues in videos where rolling text at the bottom of the screen intersects with subtitles. It utilizes easyocr to detect text in frames and creates bounding boxes. The program checks for the presence of subtitles in each frame and calculates the intersection between the bounding box and subtitle area.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages