Hello! This is known as "object segmentation" using edge-based features. For specifics, we can use canny, prewitt (gradients) or frequency based filters to do so, to find the CONTOURS as you seek, followed by the segmentation part.
Regarding your project:
I've few queries starting with understanding the exact scenario of the problem? If you can find some time for discussion, that would be great. Alternatively, you can attach a brief with sample input files and outputs. As in details like:
1) formats of I/O
2) environment, lighthing conditions
3) and, everything regarding the type of application (front-end backend stuff)
What is possible if you ask me?
Since you've mentioned Photoshop, I'm wondering if you want a plugin? As a standalone you can use Python script that is most convenient and compatible on every platform (i.e. Windows, Mac + others )
About me:
Am a research assistant in computer vision and machine learning lab. I taje projects that are along the lines of my work.
The current bid + deadline shall be revised post discussion. We shall even have a write-up/timeline to track progress and milestone payments.
Kindly leave your queries etc even if am not online. I shall revert back ASAP.
Thank you! Have a great day!