Synchronise two videos for comparison using images matching

Hi!

I am looking to compare two tracking data sets from two separate videos.

I need to quantify the differences between the two datasets, so for that I need to make sure that the sequences I compare start at the same frame.

I was thinking of using the retrieveImages function and compare two seconds of a video against a reference image. Then I'll consider that the fit for the reference image in the two seconds dataset, and make this my starting point.

I tried it and it worked but this is quite long, and it requires to have the videos converted to frames sequences beforehand.

Is there a way to look for one image in a video without converting it to a sequence?

I'd be open to solutions using other softwares but I don't know where to look.

Thank you for your help!

Marie

Answer 1

[Yogesh Khurana]

I think you need to convert the video to sequence of frames for using them against some algorithm. You can use an app for batch processing of sequence of frames from the video. The following link provide more information:

https://www.mathworks.com/help/images/batch-processing-using-the-image-batch-processor-app.html

You can also read frames from a video file using readFrame, even starting from a specific time. You can also use this function in your algorithm. Refer to documentation for more information:

https://www.mathworks.com/help/matlab/ref/videoreader.readframe.html#buuzswi-1-v

[marie chevalier]

Yes, that is what I did, I am working on image sequences. My problem is that the algorithm is complicated and quite slow. I am looking for a way to speed it up.

One that could work would be to know before hand which area of the image I want to compare between two frames. This would reduce the amount of features to look for and therefore speeding the process up.

Other tricks would be to affect the algorithm directly, and change the features it's looking at to fit my problem specifically but I am not sure this is possible.