This repository is a partly copy of my previous repository. It use machine learning techniques such as Convolutional Neural Network (CNN) to find out the areas of objects in images.
Some examples of result:
My main contribution is building up the utility scripts as well as the part of creating and adding soft-weights to the DRS module.
The full version of this work can be found at: https://github.com/LucasZhan/DRS_Lucas
The original DRS can be found at: https://github.com/qjadud1994/DRS