arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2104.13876cs.CV
收藏

PDNet:通过预测解耦实现更好的一级目标检测

PDNet: Toward Better One-Stage Object Detection With Prediction Decoupling

Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);University of Chinese Academy of Sciences(中国科学院大学);The Chinese University of Hong Kong(香港中文大学);PeopleAI, Inc.(PeopleAI公司);CAS Center for Excellence in Brain Science and Intelligence Technology(中国科学院脑科学与智能技术卓越创新中心)

Li Yang, Yan Xu, Shaoru Wang, Chunfeng Yuan, Ziqi Zhang, Bing Li, Weiming Hu

摘要:

最近的一阶段对象检测器遵循逐像素预测方法,该方法预测来自每个单个网格位置的对象类别分数和边界位置。然而,推断不同目标的最合适位置,即,对象类别和边界通常是不同的。因此,从同一网格位置预测所有这些目标可能导致次优结果。在本文中,我们分析了合适的推理位置的对象类别和边界,并提出了一个预测目标解耦检测器命名为PDNet建立一个更灵活的检测范式。我们的PDNet与预测解耦机制分别在不同的位置编码不同的目标。一个可学习的预测收集模块被设计成具有两组动态点,即,动态边界点和语义点,以收集和聚合来自有利区域的预测用于定位和分类。我们采用两步策略来学习这些动态点的位置,首先为不同的目标估计先验位置,然后网络进一步预测位置的残余偏移,从而更好地感知对象属性。在MS COCO基准上的大量实验证明了该方法的有效性和效率。以单个ResNeXt-64 x4 d-101-DCN作为主干,我们的检测器在单尺度测试中达到50.1 AP,在相同的实验设置下,该检测器的性能明显优于最先进的方法。此外,我们的检测器作为一个单阶段框架是非常高效的。我们的代码在https://github.com/yangli18/PDNet上公开。

英文摘要:

Recent one-stage object detectors follow a per-pixel prediction approach that predicts both the object category scores and boundary positions from every single grid location. However, the most suitable positions for inferring different targets, i.e., the object category and boundaries, are generally different. Predicting all these targets from the same grid location thus may lead to sub-optimal results. In this paper, we analyze the suitable inference positions for object category and boundaries, and propose a prediction-target-decoupled detector named PDNet to establish a more flexible detection paradigm. Our PDNet with the prediction decoupling mechanism encodes different targets separately in different locations. A learnable prediction collection module is devised with two sets of dynamic points, i.e., dynamic boundary points and semantic points, to collect and aggregate the predictions from the favorable regions for localization and classification. We adopt a two-step strategy to learn these dynamic point positions, where the prior positions are estimated for different targets first, and the network further predicts residual offsets to the positions with better perceptions of the object properties. Extensive experiments on the MS COCO benchmark demonstrate the effectiveness and efficiency of our method. With a single ResNeXt-64x4d-101-DCN as the backbone, our detector achieves 50.1 AP with single-scale testing, which outperforms the state-of-the-art methods by an appreciable margin under the same experimental settings.Moreover, our detector is highly efficient as a one-stage framework. Our code is public at https://github.com/yangli18/PDNet.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑