[Paper Review] Shape-IoU: More Accurate Metric considering Bounding Box Shape and Scale

Hao Zhang, Shuaijie Zhang|arXiv (Cornell University)|Dec 29, 2023

Advanced Image and Video Retrieval Techniques71 citations

TL;DR

This document provides a guide to using IEEEtran LaTeX templates, outlining design intents, usage, and best practices for IEEE submissions.

ABSTRACT

As an important component of the detector localization branch, bounding box regression loss plays a significant role in object detection tasks. The existing bounding box regression methods usually consider the geometric relationship between the GT box and the predicted box, and calculate the loss by using the relative position and shape of the bounding boxes, while ignoring the influence of inherent properties such as the shape and scale of the bounding boxes on bounding box regression. In order to make up for the shortcomings of existing research, this article proposes a bounding box regression method that focuses on the shape and scale of the bounding box itself. Firstly, we analyzed the regression characteristics of the bounding boxes and found that the shape and scale factors of the bounding boxes themselves will have an impact on the regression results. Based on the above conclusions, we propose the Shape IoU method, which can calculate the loss by focusing on the shape and scale of the bounding box itself, thereby making the bounding box regression more accurate. Finally, we validated our method through a large number of comparative experiments, which showed that our method can effectively improve detection performance and outperform existing methods, achieving state-of-the-art performance in different detection tasks.Code is available at https://github.com/malagoutou/Shape-IoU

Motivation & Objective

• Explain the purpose and design intent of IEEEtran LaTeX templates.
• Describe how to select appropriate documentclass options for different IEEE publication types.
• Show how to structure front matter, sections, figures, tables, and references.
• Explain the workflow from LaTeX to XML for IEEE production and IEEEXplore conversion.

Proposed method

• Describe documentclass options for journal, conference, compsoc, and technote formats.
• Present common front matter elements (title, authors, running heads, abstract, keywords).
• Provide examples for figures, tables, citations, and math formatting in IEEE style.
• Outline the production workflow from LaTeX to XML and final IEEEXplore output.
• Offer a final checklist to ensure compatibility with IEEE requirements.

Figure 1: This is the caption for one fig.

Experimental results

Research questions

• RQ1 What documentclass options are recommended for each IEEE publication type?
• RQ2 How should front matter and running heads be coded in IEEEtran templates?
• RQ3 What are the standard practices for figures, tables, equations, and references to meet IEEE guidelines?
• RQ4 What production steps convert LaTeX sources to final IEEE Xplore PDFs and HTML?

Key findings

• The templates are intended to approximate final page length and appearance, not produce final print quality.
• The design supports easy conversion to XML for IEEE outsource vendors and IEEEXplore production.
• The templates include sample files and guidelines for various IEEE publication types and collaborations.
• The document lists recommended packages and common LaTeX practices to ensure compatibility.
• A final checklist helps authors verify numbering, formatting, graphics, and references before submission.