Tse-Yu Pan

Tse-Yu Pan

National Taiwan University of Science and Technology
3 previous affiliations
3Machine learning papers tracked by Lacuna
Is this you?

Research Overview

Tse-Yu Pan focuses on multimedia systems, sports analytics, and virtual reality applications. The available evidence centers entirely on collaborative work published in 2022, highlighting contributions to computer vision techniques for basketball analysis and immersive environments for athletic training and 3D content creation. Recent efforts explore the intersection of artificial intelligence and interactive virtual spaces.

Research Trajectory

The available corpus for Tse-Yu Pan is highly concentrated, consisting of three collaborative papers published in 2022 at the ACM Multimedia conference. This work spans two primary domains: sports analytics through computer vision, and virtual reality (VR) systems for training and content creation. Across these projects, Pan appears as a second or third author, indicating active participation in developing applied multimedia systems rather than serving as the primary principal investigator.

In the domain of sports analytics, Pan contributed as a second author to ScoreActuary: Hoop-Centric Trajectory-Aware Network for Fine-Grained Basketball Shot Analysis. This research proposed a system to evaluate basketball shot events—such as shot type, result, and ball status—in real-time. The approach utilized a customized object detector and a trajectory-aware network to process basketball video feeds, specifically addressing the challenge of maintaining analytical accuracy across arbitrary camera movements.

Parallel to video-based analytics, Pan also explored the application of multimedia systems to athletic training. As a second author on BetterSight: Immersive Vision Training for Basketball Players, Pan helped develop an immersive VR training system designed to improve players' visual awareness. Unlike traditional vision training, this system required users to physically dribble a basketball while tracking multiple moving objects in a simulated court environment, bridging the gap between virtual cognitive training and physical motor skills.

Expanding beyond sports applications, Pan participated as a third author in the development of VR authoring tools. The paper GetWild: A VR Editing System with AI-Generated 3D Object and Terrain introduced a system that leverages artificial intelligence to streamline 3D environment creation. By allowing users to generate rough 3D models from 2D images or simple sketches, the system aimed to make 3D content creation more intuitive for junior artists, enabling them to edit AI-generated prototypes directly within a virtual space.

Current Directions

Based on the most recent publications in the evidence, all dating to 2022, Pan's research investigates the integration of artificial intelligence with interactive multimedia systems. A central theme across these projects is the translation of complex physical or spatial tasks—such as analyzing a basketball's trajectory, training athletic visual awareness, or sculpting 3D terrain—into computationally manageable frameworks. Open problems addressed in this recent work include improving the real-time robustness of computer vision models against dynamic camera movements and enhancing the usability of generative AI tools within immersive virtual reality environments.

Profile updated

See something wrong? Report a mistake