Výsledky vyhledávání

Report

Boosting Object Detection with Zero-Shot Day-Night Domain Adaptation

Autor: Du, Zhipeng, Shi, Miaojing, Deng, Jiankang

Detecting objects in low-light scenarios presents a persistent challenge, as detectors trained on well-lit data exhibit significant performance degradation on low-light data due to low visibility. Previous methods mitigate this issue by exploring ima

Externí odkaz: http://arxiv.org/abs/2312.01220

Zobrazit plný text záznamu

Report

Domain-General Crowd Counting in Unseen Scenarios

Autor: Du, Zhipeng, Deng, Jiankang, Shi, Miaojing

Domain shift across crowd data severely hinders crowd counting models to generalize to unseen scenarios. Although domain adaptive crowd counting approaches close this gap to a certain extent, they are still dependent on the target domain data to adap

Externí odkaz: http://arxiv.org/abs/2212.02573

Zobrazit plný text záznamu

Report

Design of the PMT underwater cascade implosion protection system for JUNO

Autor: He, Miao, Qin, Zhonghua, Hou, Shaojing, Jing, Xiaoping, Liu, Hongbang, Ke, Zunjian, Wu, Diru, Xie, Wan, Xu, Mehang, Chen, Fang, Lu, Junguang, Heng, Yuekun, Zhang, Jiawen, Ma, Xiaoyan, Du, Zhipeng

Photomultiplier tubes (PMTs) are widely used underwater in large-scale neutrino experiments. As a hollow glass spherelike structure, implosion is unavoidable during long-term operation under large water pressure. There is a possibility of cascade imp

Externí odkaz: http://arxiv.org/abs/2209.08441

Zobrazit plný text záznamu

Report

Redesigning Multi-Scale Neural Network for Crowd Counting

Autor: Du, Zhipeng, Shi, Miaojing, Deng, Jiankang, Zafeiriou, Stefanos

Perspective distortions and crowd variations make crowd counting a challenging task in computer vision. To tackle it, many previous works have used multi-scale architecture in deep neural networks (DNNs). Multi-scale branches can be either directly m

Externí odkaz: http://arxiv.org/abs/2208.02894

Zobrazit plný text záznamu