Search papers, labs, and topics across Lattice.
This work was supported in part by the National Key Research and Development Program of China under grant 2023YFB3209800. (Corresponding authors: Beinan Yu and Hui-Liang Shen.)Xiaokai Bai, Si-Yuan Cao, Xiaohan Zhang, Zhe Wu, and Hui-Liang Shen are with the College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou 310027, China (e-mail: shawnnnkb@zju.edu.cn, cao_siyuan@zju.edu.cn, zhangxh2023@zju.edu.cn, jeffw@zju.edu.cn, shenhl@zju.edu.cn).Lianqing Zheng is with the School of Automotive Studies, Tongji University, Shanghai 201804, China (e-mail: zhenglianqing@tongji.edu.cn).Beinan Yu is with the College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China, and also with the Jinhua Institute of Zhejiang University, Jinhua 321299, China (e-mail: mr_vernon@hotmail.com)Fang Wang and Jie Bai are with the School of Information and Electrical Engineering, Hangzhou City University, Hangzhou 310015, China, and also with the Hangzhou City University Binjiang Innovation Center, Hangzhou 310052, China (e-mail: wangf@zucc.edu.cn, baij@zucc.edu.cn)
1
0
3
By injecting 2D instance cues from camera data into the BEV space of 4D radar data, SIFormer overcomes radar's sparse geometry and significantly boosts 3D object detection accuracy.