PosNeg-Balanced Anchors with Aligned Features for Single-Shot Object Detection
This work addresses a key challenge in object detection for computer vision applications, offering an incremental improvement over existing methods.
The paper tackles the foreground-background class imbalance in single-shot object detection by introducing an Anchor Promotion Module and a Feature Alignment Module, achieving competitive performance of 40.0% mAP on MS COCO test-dev set with a speed of 28.6 fps.
We introduce a novel single-shot object detector to ease the imbalance of foreground-background class by suppressing the easy negatives while increasing the positives. To achieve this, we propose an Anchor Promotion Module (APM) which predicts the probability of each anchor as positive and adjusts their initial locations and shapes to promote both the quality and quantity of positive anchors. In addition, we design an efficient Feature Alignment Module (FAM) to extract aligned features for fitting the promoted anchors with the help of both the location and shape transformation information from the APM. We assemble the two proposed modules to the backbone of VGG-16 and ResNet-101 network with an encoder-decoder architecture. Extensive experiments on MS COCO well demonstrate our model performs competitively with alternative methods (40.0\% mAP on \textit{test-dev} set) and runs faster (28.6 \textit{fps}).