Portrait of Ying Lu

Ying Lu



I am an AI researcher at DiDi Research America, Mountain View, CA. At DiDi, I am working on car video data analysis and processing for smart driving.

Before that, I received my Ph.D. degree in Computer Science Department at University of Southern California (USC), under the supervision of Prof. Cyrus Shahabi. In addition, I received the M.Phil. degree from School of Information at Renmin University of China (RUC) in 2012, under the supervision of Prof. Jiaheng Lu, and received the B.Eng degree from Zhengzhou University in 2009.


Research Interests

I am generally interested in data management systems, location-based services, crowdsourcing and multimedia systems. In particular, I am interested in:

  • Large-scale Spatial Data Management: Indexing, query processing and optimization for big geo-data, such as geo-tagged mobile videos, drone videos, photos, texts, etc. The challenges are that 1) such data not only include location information but also include texts / documents, camera viewing orientations or visual contents; 2) in addition, the data are in large scale and keep increasing.
  • Route Planning: Finding a path on road networks by optimizing user-specified multi-preferences (e.g., attractiveness, safesty, cleanness, travel distance or time) or by restricting the travel time of the path within a budget (say 1 hour). Focus on efficient route planning algorithms on large-scale real-world road networks for interactive mobile and online navigation applications.
  • Places of Interest Detection: A large repository of user-generated videos / photos creates an opportunity for users to explore attractions or discover what is going on in a place thoroughly. For example, "what interesting events have happened in a city in the last 24 hours" The challenge is how to detect places of interest effectively and efficiently given a large number of videos / photos in a big city.
  • Computer Vision Applications with Geo-tagged Videos / Photos: Leverage the geo-information of videos / photos to improve the performance of video / image processing algorithms for computer vision applications (e.g., panorama generation, 3D model reconstruction, target tracking, video summarization).