Human Action Detection using Motion Sensor Data

This project focuses on building a machine learning pipeline to detect and classify human actions based on time-series motion or sensor data, such as accelerometer or gyroscope readings. Such a system has practical applications in health monitoring, activity recognition, fitness tracking, and gesture-based control systems.

It walks through preprocessing, feature engineering, model training, and evaluation using popular ML techniques

---

Table of Contents

• Overview
• Features
• Project Structure
• Getting Started
• Data Sources
• Usage
• Model Details
• Results & Evaluation
• Requirements

---

Overview

This repository provides a complete workflow for predicting and analyzing human actions from motion or sensor datasets using machine learning algorithms. The project is implemented entirely in Python Jupyter Notebooks for transparency and reusability.

---

Features

• End-to-End Pipeline: Data loading, preprocessing, feature engineering, model training, evaluation, and prediction.
• Supervised Learning: Implements classifiers such as Random Forest, SVM, and more.
• Sensor Data Analysis: Handles motion/sensor CSV datasets, visualizes signals, and explores statistical properties.
• Model Performance Evaluation: Plots confusion matrix, accuracy, precision, recall, and F1-score.
• Google Colab Ready: Easy to run in Colab or Jupyter.

---

Project Structure

Human-Action-Detection/
│
├── Human_Action_Detection.ipynb   # Main Jupyter Notebook for the entire ML workflow
├── data/                         # Raw and processed sensor/motion data
│   └── sensor_data.csv
├── models/                       # Saved model files (.pkl or .joblib)
├── images/                       # Plots/figures from analysis
├── README.md                     # Project documentation
└── requirements.txt              # Python dependencies

---

Getting Started

1. Clone the repository:

   git clone https://github.com/AnanyaGubba/Human-Action-Detection.git

2. Run on Google Colab or Jupyter:

   • Open Human_Action_Detection.ipynb in your Colab or local Jupyter environment.

3. Install dependencies:

   pip install -r requirements.txt

4. Prepare the dataset:

   • Place your sensor or motion data file in the data/ folder and update the notebook path if needed.

---

Data Sources

• Public datasets (links or references provided in the notebook)
• Typical data format: CSV files with time series sensor/motion readings and labeled actions

---

Usage

• Notebook: Run notebook cells in order for data processing, training, and evaluation.
• Custom Data: Replace sensor_data.csv with your own motion/sensor data.
• Model Saving/Loading: Code included for saving and reusing trained models.

---

Model Details

• Algorithms: Random Forest, Support Vector Machine (SVM), Logistic Regression, etc.
• Feature Engineering: Time-domain and frequency-domain features, normalization.
• Training: Train/Test split, cross-validation, hyperparameter tuning.
• Evaluation: Visualization of results, metrics calculation.

---

Results & Evaluation

• Confusion matrix, classification report, accuracy plots
• Example result summaries and best model selection
• Key insights noted in the notebook documentation

---

Requirements

The required Python packages are listed in requirements.txt.
To install them, run:

pip install -r requirements.txt

Typical dependencies include:

- numpy
- pandas
- scikit-learn
- matplotlib
- seaborn
- joblib

Please see the requirements.txt file for the complete list and versions.