Yitong Quan

Research with an Engineering Focus

How can one fisheye camera provide reliable surround-view perception? How can a robot retain object information through occlusion? How can lightweight detectors benefit from video context without expensive temporal modules?

I investigate these questions through model development, geometric methods, dataset preparation, quantitative evaluation, and integration on resource-constrained robotic platforms.

Background

2020-2025

Research assistant at the Department of Cognitive Systems, University of Tübingen

Camera-based perception, surround-view localization, spatial memory maps, and lightweight video object detection.

2020

Algorithmic developer at Spleenlab GmbH

LiDAR- and vision-based localization, mapping, obstacle detection, and ROS-based robotic integration.

2019

Software Developer at FERCHAU GmbH / E&K Automation

Camera and point-cloud software for industrial goods and pallet detection and positioning.

2014 - 2018

Master of Science in Embedded Systems Engineering at the University of Freiburg

2009 -2013

Bachelor of Engineering in Automotive Engineering at South China Agricultural University

Research Interests

Machine Learning
Deep Neural Networks
Computer Vision

Teaching

Assignments class: Vorlesung Mobile Roboter (Summer semester 2021)
Assignments class: Grundlagen der Robotik (Winter semester 2020)

	Publication
[1]	Xinyan Tan, Fangxing Li, Yitong Quan, Zhengqing Zuo, Xiaohai He, Weimin Zhang; Smoke and Fire Detection Based on Lightweight Depthwise Separable Convolution. In Proceedings of the International Conference on Automation, Robotics and Applications (ICARA), pp. 412–417, 2026. [IEEE / DOI]
[2]	Yitong Quan, Benjamin Kiefer, Martin Messmer, Andreas Zell; Learning-Based Distance Estimation for 360° Single-Sensor Setups. In 2025 European Conference on Mobile Robots (ECMR), pp. 1–8, 2025. [IEEE, arXiv]
[3]	Yitong Quan, Benjamin Kiefer, Martin Messmer, Andreas Zell; Lightweight Multi-Frame Integration for Robust YOLO Object Detection in Videos. In 2025 European Conference on Mobile Robots (ECMR), pp. 1–7, 2025. [IEEE, arXiv, GitHub]
[4]	Benjamin Kiefer, Yitong Quan, Andreas Zell; Approximate Supervised Object Distance Estimation on Unmanned Surface Vehicles. In 2025 IEEE International Conference on Advanced Robotics (ICAR), pp. 585–590, 2025. [IEEE / DOI, arXiv]
[5]	Benjamin Kiefer, Lojze Žust, Matej Kristan, et al. (including Yitong Quan); 3rd Workshop on Maritime Computer Vision (MaCVi) 2025: Challenge Results. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, pp. 1542–1569, 2025. [CVF, arXiv]
[6]	Yitong Quan, Benjamin Kiefer, Martin Messmer, Charan Ram Akupati, Rainer Graser, Andreas Zell; Robust Single-Cam Surround View Object Detection and Localization Using Memory Maps. In Proceedings of the International Conference on Pattern Recognition (ICPR), Springer, pp. 105–118, 2024. [Springer / DOI]
[7]	Benjamin Kiefer, Yitong Quan, Andreas Zell; Memory Maps for Video Object Detection and Tracking on UAVs. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 3040–3047, 2023. [IEEE / DOI, arXiv]
[8]	Benjamin Kiefer, Matej Kristan, Janez Perš, et al. (including Yitong Quan); 1st Workshop on Maritime Computer Vision (MaCVi) 2023: Challenge Results. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops, pp. 265–302, 2023. [CVF, arXiv]
[9]	Fabian Höflinger, Joan Bordoy, Rui Zhang, Yitong Quan, Amir Bannoura, Nikolas Simon, Leonhard Reindl, Christian Schindelhauer, Zhi Wang; Indoor Localization System Using Ultra Low-Power Radio Landmarks Based on Radio Signal Strength and Travel Time. In Sensor Networks: Revised Selected Papers of SENSORNETS 2017 and SENSORNETS 2018, Springer, Communications in Computer and Information Science, vol. 1074, pp. 98–117, 2019. [Springer / DOI]