arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.05997cs.CV
收藏

用于人类解析的质量感知网络

Quality-Aware Network for Human Parsing

Beijing University of Posts and Telecommunications(北京邮电大学);Institute of Automation Chinese Academy of Sciences(中国科学院自动化研究所);Noah’s Ark Lab Huawei Technologies(华为技术有限公司诺亚方舟实验室)

Lu Yang, Qing Song, Zhihui Wang, Zhiwei Liu, Songcen Xu, Zhihao Li

摘要:

如何评估网络输出的质量是一个重要的问题,目前在人工句法分析领域还没有有效的解决方案。为了解决这一问题,本文提出了一种基于输出概率图的统计方法来计算像素质量信息,称为像素得分。此外,质量感知模块(QAM)被提出来融合不同的质量信息,其目的是评估人类解析结果的质量。我们结合QAM与一个简洁有效的网络设计,提出质量感知网络(QANet)的人类解析。得益于QAM和QANet的优越性,我们在CIHP、MHP-v2、Pascal-Person-Part和LIP三个多人和一个单人句法分析基准测试中获得了最佳性能。在不增加训练和推理时间的情况下,QAM在多人解析任务中将AP $^\text{r}$标准提高了10个点以上。QAM可以扩展到具有良好质量估计的其他任务,例如实例分割。具体来说,QAM在COCO和LVISv1.0数据集上将Mask R-CNN提高了约1% mAP。基于提出的QAM和QANet,我们的整体系统在CVPR 2019 COCO DensePose Challenge中获得第一名,并在CVPR 2020 LIP Challenge的Track 1和2中获得第一名。代码和模型可在https://github.com/soeaver/QANet上获得。

英文摘要:

How to estimate the quality of the network output is an important issue, and currently there is no effective solution in the field of human parsing. In order to solve this problem, this work proposes a statistical method based on the output probability map to calculate the pixel quality information, which is called pixel score. In addition, the Quality-Aware Module (QAM) is proposed to fuse the different quality information, the purpose of which is to estimate the quality of human parsing results. We combine QAM with a concise and effective network design to propose Quality-Aware Network (QANet) for human parsing. Benefiting from the superiority of QAM and QANet, we achieve the best performance on three multiple and one single human parsing benchmarks, including CIHP, MHP-v2, Pascal-Person-Part and LIP. Without increasing the training and inference time, QAM improves the AP$^\text{r}$ criterion by more than 10 points in the multiple human parsing task. QAM can be extended to other tasks with good quality estimation, e.g. instance segmentation. Specifically, QAM improves Mask R-CNN by ~1% mAP on COCO and LVISv1.0 datasets. Based on the proposed QAM and QANet, our overall system wins 1st place in CVPR2019 COCO DensePose Challenge, and 1st place in Track 1 & 2 of CVPR2020 LIP Challenge. Code and models are available at https://github.com/soeaver/QANet.

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。

↑