Abstract:In the field of state-of-the-art object detection, the task of object localization is typically accomplished through a dedicated subnet that emphasizes bounding box regression. This subnet traditionally predicts the object's position by regressing the box's center position and scaling factors. Despite the widespread adoption of this approach, we have observed that the localization results often suffer from defects, leading to unsatisfactory detector performance. In this paper, we address the shortcomings of previous methods through theoretical analysis and experimental verification and present an innovative solution for precise object detection. Instead of solely focusing on the object's center and size, our approach enhances the accuracy of bounding box localization by refining the box edges based on the estimated distribution at the object's boundary. Experimental results demonstrate the potential and generalizability of our proposed method.

What problem does this paper attempt to address?

The paper primarily addresses the issue of imprecise localization in modern object detection technologies by proposing a new solution. Specifically, the paper focuses on how to improve the localization accuracy of bounding boxes in object detection. Traditional methods usually determine the position of an object by predicting the center position and scaling factor of the bounding box. Although widely used, this method has certain limitations in practical applications, leading to less accurate localization results. The main contributions of the paper can be summarized as follows: 1. **Proposed a new boundary distribution estimation method**: To improve the accuracy of bounding box localization, the authors propose a novel method that does not merely focus on the center and size of the bounding box but refines the edges of the bounding box by estimating the distribution of the object's boundaries. This method is accomplished in two steps: first, a rough boundary is predicted, and then, based on this rough boundary, a fine boundary distribution is further estimated, and these distributions are used to adjust the position of the bounding box. 2. **Improved the bounding box representation method**: The paper introduces an edge-centric bounding box representation method, which is more in line with the way humans manually annotate bounding boxes, i.e., adjusting each edge of the bounding box separately to match the actual boundaries of the object. This is different from the traditional center-and-size-based representation method. 3. **Proposed a coarse-to-fine boundary distribution estimation module**: To effectively utilize boundary features, the authors designed a coarse-to-fine boundary distribution estimation module. This module can significantly improve the accuracy and reliability of object detection. Experimental results show that this method achieves significant performance improvements on commonly used datasets (such as COCO), with an average precision (AP) increase of 2.0%, without adding extra computational costs or requiring additional annotation data. Moreover, the method demonstrates good generality and adaptability across different frameworks. In summary, this paper aims to improve localization accuracy in object detection through a novel boundary distribution estimation method, thereby enhancing overall detection performance.

Boundary Distribution Estimation for Precise Object Detection

Dive Deeper into Box for Object Detection

PBDE: an Effective Post-Processing Method Based on Box Density for Object Detection

Pbde: An Effective Method For Filtering False Positive Boxes In Object Detection

Real-time salient object detection with boundary information guidance

Beyond Local Search: Tracking Objects Everywhere with Instance-Specific Proposals

Tracking Randomly Moving Objects on Edge Box Proposals

Spatial Self-Distillation for Object Detection with Inaccurate Bounding Boxes

A Salient Object Detection Method Based on Boundary Enhancement

Joint Boundary-Aware Method and Bayesian Inference for Object Proposal

Localized and Precise Boundary Detection in 3-D Wireless Sensor Networks

A Robust and Efficient Boundary Point Detection Method by Measuring Local Direction Dispersion

Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection

CenterRepp: Predict Central Representative Point Set's Distribution for Detection

Distribution-Aware Calibration for Object Detection with Noisy Bounding Boxes

Confidence-driven Bounding Box Localization for Small Object Detection

BASNet: Boundary-Aware Salient Object Detection

Localized Algorithm for Precise Boundary Detection in 3D Wireless Networks

Out-of-distribution Object Detection through Bayesian Uncertainty Estimation

Boundary Estimation from Point Clouds: Algorithms, Guarantees and Applications

Center-Ness and Repulsion: Constraints to Improve Remote Sensing Object Detection via RepPoints