Research On Ranking Loss For One-stage Object Detection

The object detection algorithm is a key technology in computer vision,and it is used as a basic algorithm in many application scenarios and research fields.One-stage object detectors are trained by optimizing classificationloss and localization-loss simultaneously,with the former suffering much from extreme foreground-background class imbalance issue due to the large number of anchors.This paper alleviates this issue by proposing a novel framework to replace the classification task in one-stage detectors with a ranking task,and adopting the Average-Precision loss(AP-loss)for the ranking problem.Due to its non-differentiability and non-convexity,the AP-loss cannot be optimized directly.For this purpose,we develop a novel optimization algorithm,which seamlessly combines the error-driven update scheme in perceptron learning and backpropagation algorithm in deep networks.We provide in-depth analyses on the good convergence property and computational complexity of the proposed algorithm,both theoretically and empirically.Experimental results demonstrate notable improvement in addressing the imbalance issue in object detection over existing AP-loss optimization algorithms.An improved state-of-the-art performance is achieved in onestage detectors based on AP-loss over detectors using classification-losses on various standard benchmarks.The proposed framework is also highly versatile in accommodating different network architectures.