Skip to content

"Embark on a cutting-edge journey in Human Activity Recognition using a fusion of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks. This project includes model training, metric visualization, and action prediction in videos. Experience seamless interaction with a Streamlit-powered user-friendly version (at the bottom)

License

Notifications You must be signed in to change notification settings

billy-enrizky/Activity-Recognition-Project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Human Activity Recognition using CNN + LSTM Deep Learning Project

This project implements a Human Activity Recognition (HAR) system using a combination of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks. The goal is to recognize and classify human activities based on video input.

Table of Contents

Introduction

Human Activity Recognition (HAR) is a field of study that involves the identification and classification of human activities based on data such as video sequences. In this project, we use a combination of CNN and LSTM layers to build an effective HAR model.

Project Overview

The project is divided into two main parts:

  1. Part 1: CNN + LSTM Model Training

    • Train a model using CNN and LSTM layers for Human Activity Recognition.
    • Plot model loss and accuracy curves.
    • Implement the Long-term Recurrent Convolutional Network (LRCN) approach.
  2. Part 2: Action Recognition on Videos

    • Download YouTube videos for testing.
    • Create functions for action recognition in videos.
    • Evaluate the trained model on test videos.
    • Implement a user-friendly Streamlit version.

Requirements

  • Python 3.x
  • TensorFlow
  • Keras
  • OpenCV
  • MoviePy
  • Pafy
  • Matplotlib
  • Streamlit (for Streamlit version)

Installation

  1. Clone the repository:

    git clone <repository-url>
    cd <repository-directory>
  2. Install the required dependencies:

    pip install -r requirements.txt

Usage

  1. Part 1: Model Training

    • Open and run the Jupyter notebook for Part 1.
    • Follow the instructions in the notebook to train the CNN + LSTM model.
  2. Part 2: Action Recognition on Videos

    • Open and run the Jupyter notebook for Part 2.
    • Follow the instructions in the notebook to download YouTube videos and perform action recognition.

Models

  • Two models are implemented in this project:

    1. ConvLSTM Model
    2. Long-term Recurrent Convolutional Network (LRCN) Model
  • The models are trained for Human Activity Recognition.

Evaluation

  • Model performance is evaluated on test datasets.
  • There are 6676 videos in UCF50 dataset.
  • Loss and accuracy curves are plotted for model evaluation.

Action Recognition on YouTube Videos

  • Functions are provided to download YouTube videos for testing.
  • Action recognition is performed on these videos using the trained model.

Streamlit User-Friendly Version

  • A Streamlit version is created for easy interaction.
  • Users can recognize actions in a video with a single click.

Contributing

Contributions are welcome! If you have suggestions or find issues, please open an issue or create a pull request.

License

This project is licensed under the MIT License.


Please make sure to replace `<repository-url>`, `<repository-directory>`, and `<link-to-issues>` with the appropriate values. Additionally, include the license file (e.g., `LICENSE`) in your project directory. Adjust the structure and content as needed for your specific project.

About

"Embark on a cutting-edge journey in Human Activity Recognition using a fusion of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks. This project includes model training, metric visualization, and action prediction in videos. Experience seamless interaction with a Streamlit-powered user-friendly version (at the bottom)

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published