我的新闻

分享到:
三篇文章被AAAI21会议接收
发布者: 洪晓鹏 | 2020-12-02

团队三篇文章被人工智能顶级会议AAAI 2021接收。恭喜 马智恒、贺宇航和董松林&陶小语,恭喜各位合作的老师。

 

 

Title: Learning to Count via Unbalanced Optimal Transport. 

中文题目:基于非平衡最优传输理论的计数学习

作者:Zhiheng Ma, Xing Wei, Xiaopeng Hong, Hui Lin, Yunfeng Qiu, Yihong Gong. 

 

摘要: Counting dense crowds through computer vision technology have attracted widespread attention. Most crowd counting datasets use point annotations. In this paper, we formulate crowd counting as a measure regression problem to minimize the distance between two measures with different supports and unequal total mass. Specifically, we adopt the unbalanced optimal transport distance, which remains stable under spatial perturbations, to quantify the discrepancy between predicted density maps and point annotations. An efficient optimization algorithm based on the regularized semi-dual formulation of UOT is introduced, which alternatively learns the optimal transportation and optimizes the density regressor. The quantitative and qualitative results illustrate that our method achieves state-of-the-art counting and localization performance. The source code is available at \url{https://xxx.xxx.xxx}.
 

 

 

 

Title: Error-Aware Density Isomorphism Reconstruction for Unsupervised Cross-Domain Crowd Counting.

中文题目:面向非监督跨域人群计数的错误可查觉密度同态重构

作者:Yuhang He, Zhiheng Ma, Xing Wei, Xiaopeng Hong, Wei Ke, Yihong Gong. 

 

摘要:This paper focuses on the unsupervised domain adaptation problem for video-based crowd counting, in which we use labeled data as the source domain and unlabelled video data as target domain. It is challenging as there is a huge gap between the source and the target domain and no annotations of samples are provided in the target domain. The key is how to open up opportunities for utilizing unlabelled videos in the target domain to learn and transfer knowledge from the source domain. To tackle this problem, we propose a novel Error-aware Density Isomorphism REConstruction Network (EDIREC-Net) for cross-domain crowd counting, which transfers a pre-trained counting model to target domains using a density isomorphism reconstruction objective and models the reconstruction erroneousness by error reasoning. Specifically, as crowd flows in videos are consecutive so that density maps in adjacent frames are isomorphic, we regard the density isomorphism reconstruction error as a self-supervised signal to transfer the pre-trained counting models to different target domains. Moreover, we leverage an estimation-reconstruction consistency to monitor the density reconstruction erroneousness and suppress unreliable density reconstructions during training. Experimental results on four benchmark datasets demonstrate the superiority of the proposed method and ablation studies investigate the efficiency and robustness.

 

 

 

Title: Few-Shot Class-Incremental Learning via Relation Knowledge Distillation.

中文题目:基于关系知识蒸馏的小样本类增量学习

作者:Songlin Dong, Xiaopeng Hong, Xiaoyu Tao, Xinyuan Chang, Xing Wei, Yihong Gong. 

 

摘要:In this paper, we focus on a challenging few-shot class-incremental learning (FSCIL) problem, which requires to transfer knowledge from old tasks to new ones and solve catastrophic forgetting. To this end, we propose the exemplar relation distillation incremental learning framework to well balance old-knowledge preserving and new-knowledge adaptation. First, we construct an exemplar relation graph to represent the knowledge of the original network and gradually update as learning the new task. Then an exemplar relationloss function for discovering the relation knowledge between different classes is introduced to learn and transfer the substructure information in relation graph. A large number of experiments demonstrate that relation knowledge does exist in the exemplars and our approach outperforms other state-of-the-art class-incremental learning methods on the CIFAR100,miniImageNet, and CUB200 datasets.
 

 

News List
闻列表