How to decrease false positives and how to separate very small adjacent instances?

[ustczhouyu]

❓ Questions and Help

Help! When I train the model on a dataset containing many small objects, I encountered other difficulties. 1. The model will detect two or more small objects that are close together in the horizontal or vertical direction as one. 2. Due to the complex background of this dataset, some backgrounds are even similar to the texture of the foreground, leading to some false positives. What should I do to solve these two problems? (For example, which parameters should be modified or what kind of branch should be added?) Please help me. @mrlooi

[mrlooi]

This is a general question, and unfortunately there's no simple way to do this. You could try changing the anchor scale parameters, and write your own loss functions to further penalize false positives. Perhaps even changing the network.
Try looking at the latest CV papers, I believe the techniques used in this repo (based on rotated anchors and maskRCNN) is no longer the state of the art for general segmentation and detection.