https://github.com/abhinavsharma07/hand_gesture_recognition-deep_learning

The objective of this project is to recognize hand gestures using state-of-the-art neural networks.
https://github.com/abhinavsharma07/hand_gesture_recognition-deep_learning

cnn deep-learning keras lstm modelevaluation recurrent-neural-networks rnn tensorflow

Last synced: 5 days ago
JSON representation

The objective of this project is to recognize hand gestures using state-of-the-art neural networks.

• Host: GitHub
• URL: https://github.com/abhinavsharma07/hand_gesture_recognition-deep_learning
• Owner: AbhinavSharma07
• License: mit
• Created: 2024-08-26T13:39:00.000Z (2 months ago)
• Default Branch: main
• Last Pushed: 2024-08-26T19:32:52.000Z (2 months ago)
• Last Synced: 2024-10-10T08:22:51.857Z (29 days ago)
• Topics: cnn, deep-learning, keras, lstm, modelevaluation, recurrent-neural-networks, rnn, tensorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 17.7 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Hand Gesture Recognition - Deep Learning Project.

## 🎯 Goal

Recognize hand gestures using state-of-the-art neural networks to control smart TVs without a remote.

## 🔍 Problem Statement

As a data scientist at a home electronics company, your mission is to develop an innovative feature for smart TVs that recognizes five distinct hand gestures, enabling users to control their TV without a remote.

### 🖐️ Gestures and Commands:

- 👍 Thumbs up: Increase volume

- 👎 Thumbs down: Decrease volume

- 👈 Left swipe: Jump backwards 10 seconds

- 👉 Right swipe: Jump forward 10 seconds

- ✋ Stop: Pause the movie

Each gesture is captured as a sequence of 30 frames by a webcam mounted on the TV.

## 📊 Dataset

- Training data: Hundreds of categorized videos

- Video length: 2-3 seconds

- Frame sequence: 30 frames per video

- Recorded by: Various individuals performing gestures

- Data structure: 'train' and 'val' folders with corresponding CSV files

- Video dimensions: 360x360 or 120x160

[Download Dataset](https://drive.google.com/uc?id=1ehyrYBQ5rbQQe6yL4XbLWe3FMvuVUGiL)

## 🚀 Challenge

Train a model on the 'train' folder that performs well on the 'val' folder.

## 🛠️ Technologies Used

- Python

- TensorFlow / PyTorch

- OpenCV

- Numpy

- Pandas

## 🚀 Getting Started

1. Clone this repository

2. Download and extract the dataset

3. Install required dependencies

4. Run the Jupyter notebook or Python scripts