Regreesion and Classification Heads in Code
Could you please advise where in this project I can find the classification and regression heads? How the regression head gets bounding box coordinates as inputs to regress the object's bounding box coordinates in the next frame.
@SamihaSara
Heads are at https://github.com/wuyuebupt/doubleheadsrcnn/blob/master/maskrcnn_benchmark/modeling/roi_heads/box_head/box_head.py.
For the second question, that's a general process in object detection. Check this function: https://github.com/wuyuebupt/doubleheadsrcnn/blob/a744b4121d52935741f49d845bae7878270ea291/maskrcnn_benchmark/modeling/roi_heads/box_head/loss.py#L52
@wuyuebupt If I want to add a new field in annotation and regress it. input will be bounding box coordinates, and I have to regress this score ranging from 0 to 1. Could you please suggest which files I will need to modify for that? This operation is similar to bounding box regression. Thanks.
@SamihaSara
In my understanding, you will need to modify, data loader, head strucutre and loss.
For data loader, https://github.com/wuyuebupt/doubleheadsrcnn/blob/master/maskrcnn_benchmark/data/datasets/coco.py#L39
For head strucutre, https://github.com/wuyuebupt/doubleheadsrcnn/blob/master/maskrcnn_benchmark/modeling/roi_heads/box_head/box_head.py
For loss, https://github.com/wuyuebupt/doubleheadsrcnn/blob/master/maskrcnn_benchmark/modeling/roi_heads/box_head/inference.py https://github.com/wuyuebupt/doubleheadsrcnn/blob/master/maskrcnn_benchmark/modeling/roi_heads/box_head/loss.py
Hope these help.