Skip to content

About

Touch-free control of an information display: a 3D CNN trained on the Jester dataset recognises hand gestures from a live webcam and drives the pages through Flask-SocketIO.

Topics

Resources

Stars

0 stars

Watchers

1 watching

Forks

Latest commit

 

History

112 Commits

Folders and files

Repository files navigation

Real-time segmentation of dash-cam videos

This project segments dash-cam video frames using a U-Net or a DeepLabV3+ (ResNet-50 backbone) model trained on Indian road scenes, and renders the output as a GIF for quick review. It supports both local video files and YouTube URLs as input.

Demo

Dataset

The dataset used for this project can be found at Kaggle

Requirements

  • Python 3.9+
  • pip install -r requirements.txt

Setup

python -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install -r requirements.txt

Usage

# Segment a YouTube video
python main.py --youtube "https://youtu.be/h1u5OzTdbpc"

# Segment a local file
python main.py --file ./video/demo_video.mp4

# Optional: verbose logs
python main.py --file ./video/demo_video.mp4 --verbose

Model Performance

  • Obtained a mean_iou Score of 0.65907 on test data

Mean IoU across training epochs

Output

  • Segmented frames are written to processed_frames/.
  • Final GIF is written to images/.

About

Touch-free control of an information display: a 3D CNN trained on the Jester dataset recognises hand gestures from a live webcam and drives the pages through Flask-SocketIO.

Topics

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

Packages

Used by

Contributors

Languages