How to find marks in the image with Firebase ML Kit?

Question

I have mobile app for android. I need to process images in my app. This is example images:

This is a photo of the table with marks. I need to find these marks and their positions. Is it possible to do the test? Is it possible to do with ML Kit for Firebase?

If yes - which api should i use? I will be grateful for any advice.

Solution

I am not sure why you need ML. I would use ML if I needed to differentiate the marks in the cell between a lower letter v, uppercase letter V or an actual check mark. Your use case relates more to image processing. I would use a differential between your virgin template against the marked template and then map the detection coordinates against the initial cell's row/col value. Another possible approach is to use OCR assuming that all entries are check marks(crosses or circles but not filling the blank). Based on the use case, I would use the former.

Getting the pixel from an image, check this. This would be my initial code for the "differential" approach. This is very manual but doable. Another suggestion is to bring OpenCV to Flutter. OpenCV allows pattern detection so I have seen. You can get the corners of each cell first. Then you can do pixel processing per cell. Related to OCR, Firebase has the capability as text recognition which provides coordinates. You can review their docs here.