我的新闻

在ICCV19展示团队工作


2019-11-05

国际计算机视觉顶级会议ICCV-2019于10月27日至11月2日在韩国首尔召开。洪晓鹏老师和团队骨干魏星、马智恒一起参会并在展示了相关工作。

 

[*] Ma, Zhiheng and Wei Xing and Hong, Xiaopeng and Gong, Yihong. Bayesian Loss for Crowd Count Estimation with Point Supervision. Proceedings of the international conference on computer vision, (ICCV-19), pp. 6142-6151, 2019.

 

全文下载: http://openaccess.thecvf.com/content_ICCV_2019/papers/Ma_Bayesian_Loss_for_Crowd_Count_Estimation_With_Point_Supervision_ICCV_2019_paper.pdf

代码下载: https://github.com/ZhihengCV/Bayesian-Crowd-Counting


 

@inproceedings{BayersianLoss19,

  title={Bayesian Loss for Crowd Count Estimation with Point Supervision},

  author={Ma, Zhiheng and Wei Xing and Hong, Xiaopeng and Gong, Yihong},

  booktitle={Proceedings of the international conference on computer vision, (ICCV-19)},

  year={2019}

}

 

我们在人群计数领域提出了一种与传统方法不同的弱监督学习算法,不再依赖于对点(人头中心点)标签进行高斯模糊假设得到“伪标签”来进行模型学习,而是依靠计算模型输出的期望来得到更为准确的误差损失计算,从而进行模型学习。该工作《Bayesian Loss for Crowd Count Estimation with Point Supervision》已被国际模式识别顶级会议《International Conference on Computer Vision》录用为口头展示文章(口头录用率不足5%)。 该工作受到与会者的广泛关注。其中,国际计算机视觉著名学者美国中佛罗里达大学(UCF)计算机视觉研讨中心主任Mubarak Shah教授(IEEE,AAAS,SPIE,IAPR会士)在会场展示时给与该工作高度评价“A new crowd counting solution which is simple and elegant”。

 

 

图1  马智恒在展示工作

图2  与Mubarak Shah教授深入交流

 

此外,洪晓鹏老师本次ICCV会议还有其他三个工作以poster方式展示。

[*] Y. Xu, D. Xu, X. Hong, W. Ouyang, R. Ji, M. Xu, G. Zhao. Structured Modeling of Joint Deep Feature and Prediction Refinement for Salient Object Detection. Proceedings of the international conference on computer vision, (ICCV-19), pp. 6142-6151, 2019. [PDF]

[*] Z. Yu, W. Peng, X. Li, X. Hong, G. Zhao. Remote Heart Rate Measurement from Highly Compressed Facial Videos: an End-to-end Deep Learning Solution with Video Enhancement. Proceedings of the international conference on computer vision, (ICCV-19), pp. 6142-6151, 2019. [PDF]

[*] J. Li, R. Ji, H. Liu, X. Hong, Y. Gao, Q. Tian. Universal Perturbation Attacking against Image Retrieval. Proceedings of the international conference on computer vision, (ICCV-19), pp. 6142-6151, 2019. [PDF]