Contour Method for Differentiating between slides

[GrayFlash]

@gaulghost suggested this method in issue #10 and we will like to have more discussion about it here.

Before suggesting anything, refer to these constraints:

• We make user select screen region which needs to be captured,
• you are having two images that are taken as screenshots just milliseconds apart, so they may point to the same slide or a different one,
• The purpose is to identify if the slides are the same or different,
• Also some teachers or interviews including writing on a whiteboard on which the presenter starts writing, goes on writing and when done, clears the entire whiteboard. We only need to save the slide when the whiteboard was fully written and not when the difference between two consequent images was of a letter or word.

[gaulghost]

Hi, I am trying to firstly pin down the presentation area from the full window screen so that I am able reduce the amount of contours i need to deal with. Sources for information:- https://www.pyimagesearch.com/2017/06/19/image-difference-with-opencv-and-python/ https://www.geeksforgeeks.org/find-and-draw-contours-using-opencv-python/ Refer the above image for example we need to somehow pin down our search from three contours down to two or one. To reduce the number of countours what I tried was picking up the largest Contour (area wise) out from a number of contours available. I thought the second largest countour would be the presentation area itself since the presentation area occupies larget space... I wasn't able to test this theory out since I am getting some kind of error which i am trying to rectify, will update as soon as I correct it.

[GrayFlash]

I am trying to firstly pin down the presentation area from the full window screen so that I am able reduce the amount of contours i need to deal with.

The UI that we created earlier on Tkinter helps users select the presentation area, maybe that will be helpful for you??

I also look into the work you have shared. @shankhanil007 @D3ADSH0T25 @Apurva-tech if any one of you find it interesting feel free to get into it.

[Anshoo-Rajput]

@gaulghost getting a presentation area is not an issue. As @GrayFlash said using Tkinter UI user can specify which area is to be monitored for capturing screenshots. If somehow we can apply this method to text or handwriting of slides then comparing 2 contour images might be helpful than just comparing pixels difference in images as most of the part in slides do not change that often.

[GrayFlash]

@gaulghost I see inactivity for a very long time, it will be great if you can provide some update.

[gaulghost]

@GrayFlash Hi, when I submitted that Issue 10, I already included this feature in that one. While diffentiating beween slides, contour method was used.