Fall Detection

This repository contains code for fall detection, capable of running on both GPU (with CUDA support) and CPU. The program analyzes input data to detect falls in real time.

Prerequisites

• GPU Execution: Ensure your system has a CUDA-supported GPU and that CUDA drivers are properly installed. Refer- https://docs.nvidia.com/cuda/cuda-installation-guide-linux/
• CPU Execution: No special hardware requirements.

Running Fall Detection From Source

Installation

Clone this repository to your local machine-

git clone https://github.com/sahithchada/fall_detect.git

Install all the requirements using-

pip3 install -r requirements.txt

Downloading the model Download the weights from google drive link below and place it inside Models directory-

https://drive.google.com/drive/folders/1rZit5YD-Cvx_lhvEKrE5iPIRsNzR6Lir?usp=sharing

Running Source Code

python3 blur_face_and_detect.py

Running the Fall Detection Program From Spec File

Using GPU (with CUDA support)

If your system has a GPU with CUDA support, follow these steps to run the fall detection program:

1. Open the directory in your terminal:

   cd fall_detect_x64_package

2. Make the GPU executable file runnable (this needs to be done only once):

   chmod +x fall_detect_gpu

3. Run the GPU version of the fall detection program:

   ./fall_detect_gpu

Using CPU

If you are using a CPU, follow these steps to run the fall detection program:

1. Open the directory in your terminal:

   cd fall_detect_x64_package

2. Make the CPU executable file runnable (this needs to be done only once):

   chmod +x fall_detect_cpu

3. Run the CPU version of the fall detection program:

   ./fall_detect_cpu

Additional Information

The program will output the detection results to a separate window.