A Systematic IoU-Related Method: Beyond Simplified Regression for Better Localization
share
Four-variable-independent-regression localization losses, such as Smooth-ℓ_1 Loss, are used by default in modern detectors. Nevertheless, this kind of loss is oversimplified so that it is inconsistent with the final evaluation metric, intersection over union (IoU). Directly employing the standard IoU is also not infeasible, since the constant-zero plateau in the case of non-overlapping boxes and the non-zero gradient at the minimum may make it not trainable. Accordingly, we propose a systematic method to address these problems. Firstly, we propose a new metric, the extended IoU (EIoU), which is well-defined when two boxes are not overlapping and reduced to the standard IoU when overlapping. Secondly, we present the convexification technique (CT) to construct a loss on the basis of EIoU, which can guarantee the gradient at the minimum to be zero. Thirdly, we propose a steady optimization technique (SOT) to make the fractional EIoU loss approaching the minimum more steadily and smoothly. Fourthly, to fully exploit the capability of the EIoU based loss, we introduce an interrelated IoU-predicting head to further boost localization accuracy. With the proposed contributions, the new method incorporated into Faster R-CNN with ResNet50+FPN as the backbone yields 4.2 mAP gain on VOC2007 and 2.3 mAP gain on COCO2017 over the baseline Smooth-ℓ_1 Loss, at almost no training and inferencing computational cost. Specifically, the stricter the metric is, the more notable the gain is, improving 8.2 mAP on VOC2007 and 5.4 mAP on COCO2017 at metric AP_90.
READ FULL TEXT