top of page
  • Jane Shen

Call for Papers & Participation: CVPR 2020 Anti-UAV Workshop & Challenge

As the "Olympic Games" in the field of computer vision, CCF Class A International Conference CVPR 2020 (IEEE Conference on Computer Vision and Pattern Recognition) will be held in Seattle, USA on June 16-20, 2020. Recently, CVPR's official website announced the final results of listed proposals. Various workshops and challenges have started registration.

Among them, the 1st Anti-UAV Workshop & Challenge organized by "All-China Class" AI scholars is particularly eye-catching. The workshop is conducted by Dr. Zhao Jian, assistant researcher at Institute of North Electronic Equipment, Jane Shen Shengmei, chief scientist at Pensees, Dr. Wang Qiang & researcher Dr. Xing Junliang, & associate researcher Dr. Zhu Guibo at Institute of Automation, Chinese Academy of Sciences, Hu Kai, postgraduate student from Tianjin University, Dr. Hong Xiaopeng, distinguished researcher at Xi'an Jiaotong University, Dr. Guo Yandong, chief scientist at XPeng, Dr. Zhang Tianzhu, professor at University of Science and Technology of China, Dr. Feng Jiashi, assistant professor of National University of Singapore, Dr. Guo Guodong, director of Institute of Deep Learning at Baidu and many other scholars, and co-sponsored by Pensees, Kingsoft Cloud, and Drones. In addition to receiving related papers, it will also organize the first Anti-UAV Challenge. The challenge is mainly for visual perception and processing tasks such as detection, tracking, and recognition of drones in complex environments based on multi-modal video stream data.

In recent years, commercial small-sized drones have developed rapidly. Compared with manned aircraft, they have advantages such as small size, low cost, and strong maneuverability, able to undertake tasks that cannot be completed by manned aircraft. Thus, they have been widely used in many fields, including aerial photography, monitoring, telemetry, exploration, rescue, logistics. However, the wide use is accompanied by many unlawful invasion incidents of drones at home and abroad in recent years, which not only caused serious harm to citizens' personal privacy and safety of life and property, but also brought serious threats to security in sensitive areas such as airports, military bases, large-scale assembly sites, nuclear power plants, and government premises. Therefore, it is of great significance to conduct research on intelligent perception of low, slow and small (drone) targets in complex environments, so as to effectively detect and monitor drones.

The original intention of the CVPR 2020 Anti-UAV Workshop & Challenge is to encourage experts, scholars, teachers, and students from the fields of multi-modal small object detection, tracking, recognition and other related fields to display their scientific research results, and discuss, exchange and conceive new ideas, new solutions together, and thereby promote performance and value of relevant technologies and systems in practical application scenarios.

The agenda of CVPR 2020 Anti-UAV Workshop & Challenge lasts for 0.5 day (afternoon on June 19, 2020). Dr. Ling Haibin, professor at Stony Brook University, New York, and Dr. Yang Mingxuan, professor at the University of California, Merced will be invited to give keynote speeches. 20 academic posters and 1 oral presentation (from the best paper team) are expected. Details are as follows.


Paper Submission

Key dates

  • Closing date: March 24, 2020

  • Notification of paper review: April 8, 2020

  • Submission of final paper: April 16, 2020

  • Conference: June 19, 2020

Website of paper submission:

Topics of interest

The submissions are expected to deal with visual perception and processing tasks which include but are not limited to:

  • Applications of computer vision on UAVs

  • Strategies for searching of UAVs based on NIR and/or VIS data

  • Spectrum sensing techniques for UAVs detection

  • Localization and open-set identification of UAVs

  • Scene understanding for UAVs

  • Small/tiny object detection and tracking techniques

  • Fine-grained object recognition

  • Real-time deep learning inference

  • Infrared image and video analysis

  • Multimodal fusion techniques



Key dates

  • Submission: April 15, 2020

  • Conference: June 19, 2020

The CVPR 2020 Anti-UAV Challenge requires algorithms to deliver accurate, stable, and real-time tracking of a given UAV target in multi-modal video stream and simultaneously estimate the tracking states of the target. When the target disappears, an invisible mark of the target needs to be given.

The dataset provided by organizer consists of 160 high quality, full HD bimodal video sequences (both RGB and Thermal Infrared, 100 validation sets and 60 test sets), spanning multiple occurrences of multi-scale UAVs (3 sizes: large, medium, small; 4 models: DJI-Inspire, DJI-Phantom4, DJI-MarvicAir, DJI-MarvicPRO). The visible light and near-infrared video data are collected by a special ground-mounted camera device with an automatic turntable that can be remotely controlled by a computer. All data is labeled by professional data labelers, with the following labeling information: bounding box, attribute (big, medium, small, day, night, cloud, building, false object, speed change, hang, occlusion, scale variation) and a flag bit indicating whether a target exists in the current frame.

Relevant data sets have been opened for download. Team registration and results submission have started. For more details, please refer to:



Agenda: afternnon on June 19, 2020

  • 13:30-14:00 Opening ceremony and speech

  • 14:00-14:30 Challenge introduction and results announcement

  • 14:30-14:40 Oral presentation 1: Champion of Anti-UAV Challenge

  • 14:40-15:10 Invited speaker 1: Haibin Ling, Professor, Stony Brook University

  • 15:10-15:40 Poster presentation and tea break

  • 15:40-15:50 Oral presentation 2: Runner-up of Anti-UAV Challenge

  • 15:50-16:20 Invited speaker 2: Ming-Hsuan Yang, Professor, University of California, Merced

  • 16:20-16:30 Oral presentation 3: second runner-up of Anti-UAV Challenge

  • 16:30-16:40 Oral presentation 4: Best paper

  • 16:40-17:10 Award ceremony, conclusion and prospect

Host members

  • Zhao Jian, assistant researcher at Institute of North Electronic Equipment

  • Jane Shen Shengmei, chief scientist and managing director at Pensees Singapore

  • Wang Qiang, doctoral student from Chinese Academy of Sciences

  • Xing Junliang, professor at Chinese Academy of Sciences

  • Hu Kai, postgraduate student from Tianjin University

  • Zhu Guibo, associate professor at Chinese Academy of Sciences

  • Hong Xiaopeng, distinguished researcher at Xi'an Jiaotong University

  • Guo Yandong, chief scientist at XPeng

  • Zhang Tianzhu, professor at University of Science and Technology of China

  • Feng Jiashi, assistant professor at National University of Singapore

  • Guo Guodong, director of Institute of Deep Learning at Baidu

To know more:

Jane Shen Shengmei, chief scientist and managing director at Pensees Singapore

Jane Shen Shengmei is a leader of AI and deep learning. She used to be the assistant director of Panasonic R&D Center Singapore, leading an algorithm research team of above 40 members with more than 300 patents. She has been working on relevant studies of AI in autonomous vehicles, electrocardiograph aided diagnosis and other fields. She joined Panasonic Singapore Lab in 1992, and had been engaged in the design and application of audio and video signal processing and compression algorithms, and later turned to research on image recognition. In 2013, she took the lead in leading the team to deep learning, and has achieved remarkable accomplishments in the fields of deep learning and computer vision. World-class achievements have been made in the fields of deep learning-based face detection and recognition, pedestrian detection and tracking, pedestrian re-identification (pedestrian ReID), vehicle recognition, autonomous driving, driver behavior detection, and mobile operated robots, which demonstrates her full-stack capability of computer vision.

In March 2019, She joined a Chinese AI company named Pensees as the chief scientist and managing director of the Singapore Research Institute. She has been committed to research on related technologies in the fields of monitoring and safety, smart city, autonomous driving, intelligent robots, and AI factory automation. In 2019, Pensees Singapore Research Institute under her lead won 13 world championships in computer vision technology, covering face recognition, pedestrian ReID, Vehicle ReID, and anomaly detection and other fields.

Dr. Zhao Jian at Institute of North Electronic Equipment

Zhao Jian (personal homepage: received his Ph.D. from National University of Singapore on April 30, 2019 under supervision of Feng Jiashi and Yan Shuicheng. He is currently an assistant researcher at Institute of North Electronic Equipment, a committee member of CSIG-BVD, and a member of IEEE, CCF, CSIG, BSIG. His research interests include AI, deep learning, pattern recognition, computer vision and multimedia analysis. Currently, he has participated in 2 key projects (ranked 3rd and 5th respectively). So far, he has published more than 30 high-level academic papers, with a highest impact factor of 17.73, in the following top international journals and conferences: T-PAMI, IJCV, T-IP, NeuroIPS, CVPR, IJCAI, ECCV, ACM MM, AAAI, BMVC, and WACV.

He won the Lee Hwee Kuan Award (Gold Award) as 1st author in PREMIA 2019, the Best Student Paper Award in ACM MM 2018, the 1st-Place Award in MS-Celeb-1M face recognition hard set/random set/low-shot learning challenges with ICCV 2017, the 2nd-Place Award in L.I.P human parsing/pose estimation challenges with CVPR 2017, the 1st-Place Award in 2017 NIST IJB-A unconstrained face verification/identification challenges. He was an invited reviewer of IJCV, T-MM, T-IFS, T-CSVT, Neurocomputing, CSSP, JVCI, NeurIPS (one of the top 30% highest-scoring reviewers of NeurIPS 2018), CVPR, ICCV, ACM MM, AAAI, ICLR, ICML, UAI, and a Co-Chair of the CVPR 2018 L.I.P Workshop and MHP Challenge (

<The End>




bottom of page