Gee-Sern (Jison) Hsu

Professor, National Taiwan University of Science and Technology (NTUST)

Joint Professor, Institute of Artificial Intelligence, Institute of Intelligent Manufacturing, and Department of Electronic Engineering.

Professional Profile

Professor Gee-Sern (Jison) Hsu has been engaged in academic research and teaching at NTUST since 2007, following extensive experience in industrial research and development.

His expertise encompasses computer vision, deep learning, generative artificial intelligence, and intelligent visual systems. His professional experience includes research and development management, project leadership, product development, academic teaching, and international research collaboration.

Research Interests

His research focuses on computer vision and generative AI, with particular interests in:

• Generative image modeling and face reenactment

• Facial analysis, age estimation, and vision-language models

• Diffusion models and controllable image generation

• Industrial anomaly detection and visual representation learning

• 3D computer vision and reconstruction

Recent work includes large-pose face reenactment, facial age estimation using mixture-of-experts architectures, diffusion-based image synthesis, and zero-shot anomaly detection.

Selected Recent Publications

• Pose Adapted Shape Learning for Large-Pose Face Reenactment. CVPR 2024.

• Dynamic Age Estimation via Mixture of Experts: Bridging Semantic and Structural Models. MVA 2025.

• AgeMoVLE: Age Estimation with Mixture of VL Experts. CVPR Workshops 2026.

• Diffusion-to-Detection (D2D): A Training-Free Framework for Zero-Shot Anomaly Detection. Neurocomputing 2026.

For a complete publication record and further information about ongoing research, please visit the Google Scholar profile and Artificial Vision Laboratory website.