Download Project Document/Synopsis
Abnormal activity detection plays a very important role in surveillance applications. To capture the abnormal activity automatically capturing the video needs to be implemented.
Our Intelligent Video Surveillance System detects an abnormality in a video using deep learning techniques. The activities can also be detected in real-time, and these video frames will later save as an image in the system for the user to view.
Activity recognition techniques use a model with computationally complex classifiers, creating hurdles in obtaining quick responses for abnormal activity. The system will detect abnormal activity with humans in the surveillance stream using an effective Spatial autoencoder.
This system detects if there are any abnormal activities like violence or theft in the video or in real-time. The user has to upload a video for detection and the video can be uploaded from the device. The system can also detect activities in real-time. With just a click of a button from the user’s device camera, it will start detection. So, when any abnormality is detected, the video frames will be saved as an image.
The front-end involves Html, CSS, and JavaScript and the back-end involves Python. The framework used is Django and the database is MySQL. Here, we have used Avenue Dataset for Abnormal Event Detection and UCSD Anomaly Detection. And the Spatial Autoencoder model is been implemented.
Advantages
- It is easy to maintain.
- It is user-friendly.
- The system detects an abnormality in a video easily.
- It can also detect real-time activities.
- When abnormalities are found, the video frames will be saved as images.