Yujiao Shi

I am an Assistant Professor at ShanghaiTech University. Previously, I was a Postdoc and PhD student at Australian National University (ANU), supervised by Prof. Hongdong Li. My research topics include satellite image-based localization, cross-view synthesis, 3D reconstruction.

I was a research intern at Meta Reality Lab Researchs (RLR) from Jan 2022 to Jun 2022, working with Dr. Ziyun Li on burst image super-resolution and on-sensor computing. I also did an internship at Tencent XR Lab from Jul to Oct 2022.

Email / CV / Google Scholar / Twitter / Github / Linkedin

News

Postdoc and visiting positions available, please drop me an email if you are interested.
05/2024: I am serving as an Area Chair for NeurIPS 2024
04/2024: Our workshop proposal "UAVs in Multimedia" is accepted at ACM MM 2024
01/2024: One paper accepted to ICRA 2024
10/2023: Co-organizer of workshop "UAVs in Multimedia" at ACM MM 2023
09/2023: One paper accepted to NeurIPS 2023
06/2023: One paper accepted to IROS 2023
06/2023: Tutorial speaker on cross-model camera localization at CVPR 2023

Selected Publications

	Boosting 3-DoF Ground-to-Satellite Camera Localization Accuracy via Geometry-Guided Cross-View Transformer Yujiao Shi , Fei Wu, Akhil Perincherry, Ankit Vora, and Hongdong Li ICCV, 2023 paper / code This work proposes a decoupled rotation and translation estimation method for cross-view image matching, achieving significant performance improvement.
	CVLNet: Cross-View Semantic Correspondence Learning for Video-based Camera Localization Yujiao Shi , Xin Yu, Shan Wang, and Hongdong Li ACCV, 2022 paper / code This work addresses city-scale satellite image-based camera localization by using a sequence of ground-view images.
	Beyond Cross-view Image Retrieval: Highly Accurate Vehicle Localization Using Satellite Image Yujiao Shi , Hongdong Li CVPR, 2022 paper / code We introduce a new pose optimization method to accurately pinpoint which pixel in a satellite image corresponds to the query camera location.
	Accurate 3-DoF Camera Geo-Localization via Ground-to-Satellite Image Matching Yujiao Shi , Xin Yu, Liu Liu, Dylan Campbell, Piotr Koniusz, and Hongdong Li TPAMI, 2022 paper / code We propose projective transform, which (1) compliments polar transform to achieve better coarse localization performance and (2) provides a novel handcrafted method to accurately localize query camera on its matching satellite image.
	Geometry-Guided Street-View Panorama Synthesis from Satellite Imagery Yujiao Shi , Dylan Campbell, Xin Yu, and Hongdong Li TPAMI, 2022 paper / code Satellite to street-view panorama synthesis, implicit satellite image height map estimation.
	Self-Supervised Visibility Learning for Novel View Synthesis Yujiao Shi , Hongdong Li, and Xin Yu CVPR, 2021 paper / code We estimate target-view depth and source-view visibility in an end-to-end manner.
	Where am I looking at? Joint Location and Orientation Estimation by Cross-View Matching Yujiao Shi , Xin Yu, Dylan Campbell, and Hongdong Li CVPR, 2020 paper / code The first 3-DoF camera pose estimation framework via ground-to-satellite image matching.
	Optimal Feature Transport for Cross-View Image Geo-Localization Yujiao Shi , Xin Yu, Liu Liu, Tong Zhang, and Hongdong Li AAAI, 2020 paper / code Motivated by optimal transport, we invent a cross-view feature transport module to bridge the cross-view domain gap.
	Spatial-Aware Feature Aggregation for Cross-View Image based Geo-Localization Yujiao Shi , Liu Liu, Xin Yu, and Hongdong Li NeurIPS, 2019 paper / code A polar transform is introduced to bridge the ground-and-satellite domain gap, which significantly boosts the state-of-the-art cross-view localization performance.