Towards Accurate One-Stage Object Detection with AP-Loss

Abstract

One-stage object detectors are trained by optimizing classification-loss 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 verify good convergence property of the proposed algorithm theoretically and empirically. Experimental results demonstrate notable performance improvement in state-of-the-art one-stage detectors based on AP-loss over different kinds of classification-losses on various benchmarks, without changing the network architectures. Code is available at https://github.com/cccorn/AP-loss.

arXiv comment13 pages, 7 figures, 4 tables, main paper + supplementary material, accepted to CVPR 2019

References in corpus (6)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• DSSD : Deconvolutional Single Shot Detector
• Tiny-DSOD: Lightweight Object Detection for Resource-Restricted Usages
• DSOD: Learning Deeply Supervised Object Detectors from Scratch
• Training Deep Neural Networks via Direct Loss Minimization

Cited by in corpus (18)

• A Survey of Deep Learning-based Object Detection
• Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection
• AP-Loss for Accurate One-Stage Object Detection
• Optimizing Rank-based Metrics with Blackbox Differentiation
• Circle Representation for Medical Object Detection
• Dice Loss for Data-imbalanced NLP Tasks
• Improving Multispectral Pedestrian Detection by Addressing Modality Imbalance Problems
• Loss Function Discovery for Object Detection via Convergence-Simulation Driven Search
• CircleNet: Reciprocating Feature Adaptation for Robust Pedestrian Detection
• Stochastic Optimization of Areas Under Precision-Recall Curves with Provable Convergence
• DR Loss: Improving Object Detection by Distributional Ranking
• Adaptive Class Suppression Loss for Long-Tail Object Detection
• Cheaper Pre-training Lunch: An Efficient Paradigm for Object Detection
• Foreground-Background Imbalance Problem in Deep Object Detectors: A Review
• Deep AUC Maximization for Medical Image Classification: Challenges and Opportunities
• Variational Pedestrian Detection
• Addressing Class Imbalance in Scene Graph Parsing by Learning to Contrast and Score
• Revisiting the Loss Weight Adjustment in Object Detection