Scope Head for Accurate Localization in Object Detection

TL;DR

ScopeNet introduces a novel object detection approach that models location anchors as dependent relationships, combining the flexibility of anchor-free methods with precise localization, and achieves state-of-the-art results on COCO.

Contribution

The paper proposes ScopeNet, which models location anchors as mutually dependent, employs a coarse-to-fine localization strategy, and integrates an anchor selection score for improved detection confidence.

Findings

Achieves state-of-the-art results on COCO dataset.

Combines anchor-free flexibility with precise localization.

Improves confidence estimation by integrating classification and anchor scores.

Abstract

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