Scope Head for Accurate Localization in Object Detection
This addresses localization accuracy issues in object detection for computer vision applications, representing an incremental improvement over existing methods.
The paper tackles the design difficulty of hand-crafted anchors and learning complexity in object detection by proposing ScopeNet, which models anchors as mutually dependent relationships and uses a coarse-to-fine strategy, achieving state-of-the-art results on COCO.
Existing anchor-based and anchor-free object detectors in multi-stage or one-stage pipelines have achieved very promising detection performance. However, they still encounter the design difficulty in hand-crafted 2D anchor definition and the learning complexity in 1D direct location regression. To tackle these issues, in this paper, we propose a novel detector coined as ScopeNet, which models anchors of each location as a mutually dependent relationship. This approach quantizes the prediction space and employs a coarse-to-fine strategy for localization. It achieves superior flexibility as in the regression based anchor-free methods, while produces more precise prediction. Besides, an inherit anchor selection score is learned to indicate the localization quality of the detection result, and we propose to better represent the confidence of a detection box by combining the category-classification score and the anchor-selection score. With our concise and effective design, the proposed ScopeNet achieves state-of-the-art results on COCO