期刊
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
卷 45, 期 4, 页码 4932-4944出版社
IEEE COMPUTER SOC
DOI: 10.1109/TPAMI.2022.3191753
关键词
Detectors; Object detection; Feature extraction; Image edge detection; Decoding; Encoding; Training; multi-oriented object; discontinuity; aerial image
This article proposes a method to solve the discontinuity problem in multi-oriented object detection by encoding the object with double horizontal rectangles (DHRec). By arranging the coordinates of the four vertices in left-right and top-down order, the uniqueness of the encoding is ensured. The method uses area ratios to guide the decoding of the object. Experimental results show that the proposed method can accurately detect objects of arbitrary orientation and outperforms existing methods.
Most existing methods adopt the quadrilateral or rotated rectangle representation to detect multi-oriented objects. Yet, the same oriented object may correspond to several different representations, due to different vertex ordering, or angular periodicity and edge exchangeability. To ensure the uniqueness of the representation, some engineered rules are usually added. This makes these methods suffer from discontinuity problem, resulting in degraded performance for objects around some orientation. In this article, we propose to encode the multi-oriented object with double horizontal rectangles (DHRec) to solve the discontinuity problem. Specifically, for an oriented object, we arrange the horizontal and vertical coordinates of its four vertices in left-right and top-down order, respectively. The first (resp. second) horizontal box is given by two diagonal points with smallest (resp. second) and third (resp. largest) coordinates in both horizontal and vertical dimensions. We then regress three factors given by area ratios between different regions, helping to guide the oriented object decoding from the predicted DHRec. Inherited from the uniqueness of horizontal rectangle representation, the proposed method is free of discontinuity issue, and can accurately detect objects of arbitrary orientation. Extensive experimental results show that the proposed method significantly improves the existing baseline representation, and outperforms state-of-the-art methods. The code is available at: https://github.com/lightbillow/DHRec.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据