search

sgrvinod / a-PyTorch-Tutorial-to-Object-Detection

SSD: Single Shot MultiBox Detector | a PyTorch Tutorial to Object Detection
MIT License
3.04k stars 718 forks source link

Always the same bounding box predicted #34

Closed AntoineMath closed 4 years ago

AntoineMath commented 4 years ago

I modified your code to use 2 channels depth-thermal image. I have 800 training data of different body positions. But after a decent training, the predicted bbox are the same, no matter the body position in the picture. Exact same coordinates, exact same scores. (I use top_k=1 since I have only 1 object per image) Capture du 2019-12-12 14-40-56 Same with top_k = 200 : Capture du 2019-12-12 14-47-11 Capture du 2019-12-12 14-41-49

I do not use data augmentation, but I think this is another problem. It's like the loss was trapped in a local minimum, but idk. Did you have the same problem ?

EDIT: I figure out my predictions are based on the number of each body position images : left, dataset : 56 standing, 53 sitting, 73 lying down right, dataset : 56 standing, 108 sitting, 73 lying down detection_thermique

this is just a small dataset to perform tests... but I dont understand why the loss is stucked here...

AntoineMath commented 4 years ago

Solved : I did not used data augmentation. And at the moment I tried to use it, it worked. I'm surprised that it had a so huge impact.