How to detect multiple objects in a video file and label these objects with bounding boxes.

Hi, I want to know how I can create my own object detection using existing convolution neural networks like alexnet and have this applied to a video file that shows the bounding boxes of those objects.

Where are good places and sources so I can understand how to do this?

Answer 1

[Image Analyst]

Watch Joe Hicklin's demo video, that does exactly that: MathWorks Deep Learning in 11 lines of code

[Jack Zimmerman]

This is object classification.

Is this not different to multiple object detection within a a video?

If not can you give me an example?

[Image Analyst]

Well if all you need to do is to detect it, rather than figure out what it is, then it might be simpler. It might not even need a CNN. You might be able to just get by with simple thresholding. Who knows, because you dind't upload an image.

[Jack Zimmerman]

Like I said, any object would be fine, I simply want to understand how to get to the point that I can pass a video file and be able to detect multiple objects and their bounding boxes using pre-trained CNN's like alexnet.

Simple classification isnt enough, something like the documentation seen here:

https://uk.mathworks.com/help/vision/ref/trainrcnnobjectdetector.html

However, without the requirement of the ROI of the image, that can create its own bounding boxes.