About
I build generative AI systems that run in real time on real devices. At Samsung Research America I was a core contributor to Likeness (beta) — the photorealistic avatar system on Samsung Galaxy XR — working across 3D Gaussian Splatting avatar reconstruction, audio-driven facial animation, and on-device inference optimization.
I received my Ph.D. in Electrical Engineering from USC, advised by Prof. C.-C. Jay Kuo, on efficient (“green”) learning for deepfake detection and media forensics — 523+ citations, including CVPR and ICME (Oral).
Nights and weekends, I ship consumer AI products of my own.
Highlights
- Likeness on Galaxy XR — photorealistic avatars from a phone selfie scan, driven live by headset eye- and face-tracking for video calls in any app.Samsung × Google Android XR · launched Oct 2025
- Ultraland — real-time 1v1 competitive English-learning game. 100K+ users and ~10K MAU within six months, across 112 countries.Founder & solo developer
- Calendor — an end-to-end agentic AI assistant for iOS: LLM tool use across calendar, tasks, notes, and email, with voice + image input.PalmLab AI · founder & solo developer
- DefakeHop — lightweight, high-performance deepfake detection.ICME 2021 Oral · 169 citations
Selected Publications
H.-S. Chen et al., “DefakeHop: A Light-weight High-performance Deepfake Detector,” IEEE ICME 2021 (Oral). 169 citations
Y.-T. Hu, H.-S. Chen, K. Hui, J.-B. Huang, A. Schwing, “SAIL-VOS: Semantic Amodal Instance Level Video Object Segmentation,” IEEE/CVF CVPR 2019. 136 citations
H.-S. Chen, S. Hu, S. You, C.-C. J. Kuo, “DefakeHop++: An Enhanced Lightweight Deepfake Detector,” APSIPA TSIP 2023. 42 citations
H.-S. Chen, K. Zhang, S. Hu, S. You, C.-C. J. Kuo, “Geo-DefakeHop: High-Performance Geographic Fake Image Detection,” APSIPA TSIP 2023. 36 citations
H.-S. Chen, Y. Zhu, S. You, A. M. Madni, C.-C. J. Kuo, “GreenCOD: A Green Camouflaged Object Detection Method,” APSIPA TSIP 2024.
H.-S. Chen, Y. Zhu, C.-A. Yu, R. Salloum, C.-C. J. Kuo, “GIFT: A Green Image Forgery Tracker,” J. Visual Communication and Image Representation 2026.
Full list on Google Scholar.