A General Gaussian Heatmap Label Assignment for Arbitrary-Oriented Object Detection
Author(s) | Huang, Zhanchao | |
Author(s) | Li, Wei | |
Author(s) | Xia, Xiang-Gen | |
Author(s) | Tao, Ran | |
Date Accessioned | 2022-03-28T19:00:54Z | |
Date Available | 2022-03-28T19:00:54Z | |
Publication Date | 2022-02-09 | |
Description | Copyright 2022 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. This article was originally published in IEEE Transactions on Image Processing. The version of record is available at: https://doi.org/10.1109/TIP.2022.3148874 | en_US |
Abstract | Recently, many arbitrary-oriented object detection (AOOD) methods have been proposed and attracted widespread attention in many fields. However, most of them are based on anchor-boxes or standard Gaussian heatmaps. Such label assignment strategy may not only fail to reflect the shape and direction characteristics of arbitrary-oriented objects, but also have high parameter-tuning efforts. In this paper, a novel AOOD method called General Gaussian Heatmap Label Assignment (GGHL) is proposed. Specifically, an anchor-free object-adaptation label assignment (OLA) strategy is presented to define the positive candidates based on two-dimensional (2D) oriented Gaussian heatmaps, which reflect the shape and direction features of arbitrary-oriented objects. Based on OLA, an oriented-bounding-box (OBB) representation component (ORC) is developed to indicate OBBs and adjust the Gaussian center prior weights to fit the characteristics of different objects adaptively through neural network learning. Moreover, a joint-optimization loss (JOL) with area normalization and dynamic confidence weighting is designed to refine the misalign optimal results of different subtasks. Extensive experiments on public datasets demonstrate that the proposed GGHL improves the AOOD performance with low parameter-tuning and time costs. Furthermore, it is generally applicable to most AOOD methods to improve their performance including lightweight models on embedded platforms. | en_US |
Sponsor | This work was supported in part by the National Key Research and Development Program of China under Grant 2021YFB3900502, in part by the National Natural Science Foundation of China under Grant 61922013 and Grant U1833203, in part by the Beijing Natural Science Foundation under Grant L191004 and Grant JQ20021, and in part by the Aeronautical Science Foundation of China under Grant 20200051072001. | en_US |
Citation | Z. Huang, W. Li, X. -G. Xia and R. Tao, "A General Gaussian Heatmap Label Assignment for Arbitrary-Oriented Object Detection," in IEEE Transactions on Image Processing, vol. 31, pp. 1895-1910, 2022, doi: 10.1109/TIP.2022.3148874. | en_US |
ISSN | 1941-0042 | |
URL | https://udspace.udel.edu/handle/19716/30715 | |
Language | en_US | en_US |
Publisher | IEEE Transactions on Image Processing | en_US |
Keywords | Arbitrary-oriented object | en_US |
Keywords | convolutional neural network | en_US |
Keywords | gaussian heatmap | en_US |
Keywords | object detection | en_US |
Keywords | label assignment | en_US |
Title | A General Gaussian Heatmap Label Assignment for Arbitrary-Oriented Object Detection | en_US |
Type | Article | en_US |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- A General Gaussian Heatmap Label Assignment for.pdf
- Size:
- 30.83 MB
- Format:
- Adobe Portable Document Format
- Description:
- Main article
License bundle
1 - 1 of 1
No Thumbnail Available
- Name:
- license.txt
- Size:
- 124 B
- Format:
- Item-specific license agreed upon to submission
- Description: