People-In-Region-Of-Interest

Computer Vision Project to Detect people in a selected Region. This is useful for a variety of tasks such as

Security and Surveillance: In security and surveillance, a computer vision system that detects people in selected regions can be used to monitor restricted areas, such as private properties or sensitive zones. The system can identify unauthorized individuals entering these regions, triggering real-time alerts for security personnel. It can also assist in access control, tracking who enters or exits secure facilities like data centers, government buildings, or airports. Furthermore, such a system ensures real-time monitoring of high-risk areas, providing enhanced security through automated detection and rapid response.

Retail Analytics: In the retail industry, detecting people in specific regions within a store can be leveraged for foot traffic analysis, helping businesses understand customer behavior and optimize store layouts. Retailers can track how long customers linger in particular sections of the store, which can guide product placement to maximize sales. Additionally, people detection can be applied to manage queues in checkout lines or customer service areas, alerting staff when lines become too long or when additional help is needed, ultimately improving customer satisfaction and operational efficiency.

Sports and Event Management: At large sports events or public gatherings, detecting people within specific regions can help ensure crowd safety by managing crowd density in real time. For example, during a sports game or concert, the system can detect overcrowded sections or unplanned gatherings, helping organizers to redistribute people for safety. Furthermore, this technology can be used to track the movement of players within defined zones of the field or court, providing coaches and analysts with insights into player performance and positioning.

How to Run the Program

1. Run the main.py script with the video stream input.

  python main.py

A pop-up window displaying the first frame of the video will appear.

3. Select the area within the frame where you want to detect people.

4. Once you've selected the region, press Enter. The YOLOv8 model will process the selected area and detect people within it.

5. Custom Object Detection:
   By default, the model detects people (class 0). If you'd like to detect a different object instead, open detect.py and modify the classes = 0 line to the desired class number.

Want to Use Different Weights?

If you'd like to use a different set of weights, you can change the weight file in the detect.py script.

1. Open detect.py.
2. Locate the line where the weights are specified (typically weights = 'path/to/weights').
3. Replace the path with the desired weight file location.

Make sure the weight file is compatible with the model architecture you're using.