MATLAB for object tracking

Hello! I am looking for help urgently. I am working in a short project with matlab about computer vision. I have to do two object tracking methods and I am struggling a lot. I would really appreciate any help please, I am a bit lost and the project deadline is soon.

I have got a folder with different frames of a video showing a motorcycle chase from what seems to be an helicopter. So what I need to do is to pick the motorcycle from the first frame and use it to track the motorcycle in the other frames. I have already used SIFT and SURF features, so I am just asking for another method implementation. I would really appreciate if someone could provide me a working code, I am struggling a lot and I really need help. Please, I hope you can help me.

Here is the link to the frames, it is called MotorcycleChase -> https://drive.google.com/embeddedfolderviewid=1zLRwARzIfqadp8qVHiC4Y_T5U0YMQxeE#list

Answer 1

[Subash Mylraj]

Hi!

I understand that you would like to track an object across multiple images (frames from a video).

In order to track an object, you would need to first detect it. As you mentioned, you can use feature extraction algorithms like SIFT and SIRF. You can also try out object detection based neural networks like the YOLO category of networks.

Once you have the object detections, you can track them using various trackers available in MATLAB like JPDA, GNN, TOMHT, etc.

Here is a MATLAB example you can go through that tracks vehicles from a video feed:

https://in.mathworks.com/help/driving/ug/multiple-object-tracking-tutorial.html

For more information on YOLO, JPDA, GNN and TOMHT, refer the following documentation links:

• https://in.mathworks.com/help/vision/ug/getting-started-with-yolo-v3.html?searchHighlight=yolo&s_tid=srchtitle_support_results_4_yolo
• https://in.mathworks.com/help/fusion/ref/trackergnn-system-object.html
• https://in.mathworks.com/help/fusion/ref/trackerjpda-system-object.html
• https://in.mathworks.com/help/fusion/ref/trackertomht-system-object.html

Hope this helps!