Awesome

Fall-Detection

Detect humans falling on the ground from a CCTV camera feed and limit the physical damage to the person by alerting the hospital authorities.

Introduction

Elderly people tend to slip and fall in homes and remain unattended for a long time. If the affected person is not treated immediately, serious health issues including brain injuries occur. The proposed solution automatically detects humans falling, through cameras installed in homes, and sends notification to family members and hospital authorities.

Requirements

MATLAB R2013a

Usage

####Installation

1. Download zip
2. Unzip
3. Open falldetection.m in MATLAB

####MATLAB command line

falldetection <video_name>;

####Example

falldetection video.mp4;

Output

###Working

###Original video frame

###Human segemented image

###Motion History Image (MHI)

###Human shape

###Fall Detected Notification

YouTube Demo Video Link

Demo of the project

<a href="http://www.youtube.com/watch?feature=player_embedded&v=LdoLniUSOaA " target="_blank"><img src="http://img.youtube.com/vi/LdoLniUSOaA/0.jpg" alt="Youtube Video: Human Fall Detection" width="240" height="180" border="10" /></a>

Flowchart

Note

• Live transmission from the camera to the software is not supported. Instead, the data from the camera is first stored and then processed.
• The video feed should be saved in '..\Fall-Detection\Video' folder.
• The notification is sent to the user through an app installed on the phone. The apk file is not included in this repository.

Reference Paper

C.Rougier,J.MeunierFall,A.Arnaud and J.Rousseau. Detection from Human Shape and Motion History using Video Surveillance. Proceedings of the 21st International Conference on Advanced Information Networking and Applications Workshops,2007.

License