A Scale Balanced Loss for Bounding Box Regression
Degang Sun, Yang Yang, Min Li, Jian Yang, Bo Meng, Ruwen Bai, Linghan Li, Junxing Ren · IEEE Access · 2020
Object detectors typically use bounding box regressors to improve the accuracy of object localization. Currently, the two types of bounding box regression loss are ℓn-norm-based and intersection over union (IoU)-based. However, we found that these two types of losses have their drawbacks. First, for ℓn-norm-based loss, large-scale objects are more likely to obtain a larger penalty than the smaller ones when calculating localization errors, which will cause regression loss imbalance. Second, ℓn-norm-based loss has symmetry so that when the predicted bounding boxes are in some unique symmetrical relationships (i.e., Symmetric Trap), the regression loss remains unchanged. Third, for IoU-based loss, the overlap area and the union area do not change as the shape or relative position of two bounding boxes changes in some cases(i.e., Area Maze). To address these problems, we propose the scale balanced loss(LSB), which is asymmetric, position-sensitive, and scale-invariant. First, in order to obtain the property of scale invariance, it is designed as a fraction to eliminate the scale information contained in the numerator and denominator. Second, by incorporating the Euclidean distance between different corner points instead of the area, LSBis sensitive to the changes of coordinates of any corner point, so as to solve the area maze problem. Finally, by incorporating the diagonals of the overlap and the smallest enclosing rectangle, this fraction is no longer symmetric, thus solving the symmetry trap problem. To validate the proposed algorithm, we have replaced the ℓn-norm-based loss of YOLOv3 and SSD with LGIoUand LSBand evaluate their performance on Pascal Visual Object Classes and Microsoft Common Objects in Context benchmarks. The final results show that LSBhas improved their average precisions at different IoU thresholds and scales. We envision that this regression loss can also improve the performance of other visual tasks.