search

valeoai / LOST

Pytorch implementation of LOST unsupervised object discovery method
Other
234 stars 26 forks source link

corLoc evaluation #8

Closed Mirsadeghi closed 2 years ago

Mirsadeghi commented 2 years ago

Hi @osimeoni . I am suspicious about the corLoc evaluation part in the code. The corLoc for each image is true whenever one of the ground truth objects is hit! https://github.com/valeoai/LOST/blob/2b678aca89c18aa79c56ec3f6d4a0b979a91608d/main_lost.py#L311 What about other objects? Is it right?

osimeoni commented 2 years ago

Hi @Mirsadeghi,

The corloc metric is designed for one single prediction per image. It is used to measure if the predicted bounding box corresponds to any of the objects in the image as we wrote in the paper:

Correct Localization (CorLoc) metric, i.e., the percentage of correct boxes, where a predicted box is considered correct if it has an intersection over union (IoU) score superior to 0.5 with one of the labeled object bounding boxes.

The code is therefore correct and I am closing this issue. Let me know if it is not clear still.