https://github.com/atharvapathak/hand_gesture_recognition_case_study

This project involves building a 3D Convolutional Neural Network (CNN) to correctly recognize hand gestures by a user to control a smart TV using 3D-CNN and 2D-CNN + RNN/LSTM/GRU.
https://github.com/atharvapathak/hand_gesture_recognition_case_study

cnn deep-neural-networks gesture-recognition machine-learning palm-detection python pytorch resnet tensorflow video-processing

Last synced: about 6 hours ago
JSON representation

This project involves building a 3D Convolutional Neural Network (CNN) to correctly recognize hand gestures by a user to control a smart TV using 3D-CNN and 2D-CNN + RNN/LSTM/GRU.

• Host: GitHub
• URL: https://github.com/atharvapathak/hand_gesture_recognition_case_study
• Owner: atharvapathak
• Created: 2023-09-20T03:39:56.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-09-20T03:49:19.000Z (over 1 year ago)
• Last Synced: 2024-11-15T09:17:13.998Z (2 months ago)
• Topics: cnn, deep-neural-networks, gesture-recognition, machine-learning, palm-detection, python, pytorch, resnet, tensorflow, video-processing
• Language: Jupyter Notebook
• Homepage:
• Size: 17.7 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Gesture Recognition Case Study

This case study aims at the automatic interpretation of gestures in order to offer new possibilities to interact with machine and to design more natural and more intuitive interactions with computing machines.

## Table of Contents

* Problem Statement

* Objectives

* Structure of Case study

* Technologies-used

* Acknowledgements

## Problem Statement:

* The company called Home Electronics which manufactures state of art televisions and the company wants to develop a cool feature in the smart-TV that can recognise five gestures performed by the user which will help users control the TV without using the remote.

* The gestures are continuously monitered by the webcam mounted on the TV. Each gesture corresponds to a specific command:

    1) Thumbs up : Increase the volume

    2) Thumbs down : Decrease the volumn

    3) Left swipe : Jump backwards 10 seconds

    4) Right swipe : Jump forward 10 seconds

    5) Stop : Pause the movie

## Objectives:

* We are required to build a gesture recognition model using neural network which should have generators that are able to take batches of data without error and with optimized number of parameters so that the inference or prediction time should be less.

* The architecture we are going to use is:

    1) 3D Convs

    2) CNN + RNN stacked

## Structure of Case study:

* Introduction

* Problem Statement

* Objectives

* Data Preprocessing (Creating the generator)

* Model Building 

* Model Evaluation

* conclusion

## Technologies Used

- Python version 3.9.13

- numpy library version 1.20.1

- pandas library version 1.2.4

- matplotlib library version 3.3.4

- seaborn library version 0.11.1

- tensorflow library version 2.9.0

- keras library version 2.9.0

- skimage library version 0.19.2

- git version 2.33.0.windows.1

## References

1. https://github.com/amanrocks11/hand-gesture-recognition-deep-learning/blob/master/Gesture_Recognition_Final.ipynb

2. https://github.com/prateekralhan/Gesture-Recognition-Case-study-IIITB-Assignment-/blob/master/Gesture_Recognition_Case-Study_final.ipynb

3. https://smitan94.github.io/Gesture-Recognition-Neural-Network/

## Contact

Created by [https://github.com/atharvapathak] - feel free to contact me!

1. Atharva -  [email protected]