Modifying Keypoint RCNN loss functions

We tried running the KeypointRCNN model with different (hacky) loss functions, to emphasize issues where an individual point may be far off the ground truth. - max_keypoint_loss: compute the loss individually for all 5 keypoints, then take the max - hybrid_loss: average of max loss and average loss - wellplate_v8_data: non-modified version using cross entropy.

Gino Knodel

Created on January 19|Last edited on January 19

Comment

﻿
Section 1﻿
loss
loss
02k4k6k8k10kStep246810
Train (Max) Keypoint Pixel Error
Train (Max) Keypoint Pixel Error
05101520Epoch50100150200250
Val (Max) Keypoint Pixel Error
Val (Max) Keypoint Pixel Error
05101520Epoch50100150200
Run set3
﻿
﻿

Add a comment