A More Compact Object Detector Head Network with Feature Enhancement and Relational Reasoning
This work addresses the problem of inefficient feature interaction modeling in object detection for computer vision researchers, offering an incremental improvement with specific gains in compactness and performance.
The paper tackled the difficulty of reasoning about implicit feature relationships in two-stage object detectors by proposing a more compact head network (CODH) that enhances features and performs relational reasoning, resulting in a 1.3% performance boost on COCO test-dev while reducing parameters to 0.6 times those of Cascade R-CNN.
Modeling implicit feature interaction patterns is of significant importance to object detection tasks. However, in the two-stage detectors, due to the excessive use of hand-crafted components, it is very difficult to reason about the implicit relationship of the instance features. To tackle this problem, we analyze three different levels of feature interaction relationships, namely, the dependency relationship between the cropped local features and global features, the feature autocorrelation within the instance, and the cross-correlation relationship between the instances. To this end, we propose a more compact object detector head network (CODH), which can not only preserve global context information and condense the information density, but also allows instance-wise feature enhancement and relational reasoning in a larger matrix space. Without bells and whistles, our method can effectively improve the detection performance while significantly reducing the parameters of the model, e.g., with our method, the parameters of the head network is 0.6 times smaller than the state-of-the-art Cascade R-CNN, yet the performance boost is 1.3% on COCO test-dev. Without losing generality, we can also build a more lighter head network for other multi-stage detectors by assembling our method.