PeopleofSuperintelligence

Researcher

Hsuan-I Ho

Also known as Sean Ho

Hsuan-I Ho is a research scientist at NVIDIA Research and a member of NVIDIA Research Taiwan, working on human-centric computer vision and machine perception powered by generative AI. They earned a BSc in electrical engineering at National Taiwan University and a master's degree in computer science at ETH Zurich in 2021, and an early paper on summarizing first-person videos appeared at ECCV 2018. As first author, they built locally editable virtual humans and the CustomHumans dataset in 2023, the diffusion-based single-image reconstruction method SiTH in 2024, and the PHD body-fitting method in 2025. Supervised by Otmar Hilliges and Marc Pollefeys in ETH Zurich's Advanced Interactive Technologies lab, they defended a doctoral thesis on digital humans in February 2026.

Timeline

  1. 2017–2018

    First-person video summarization

    With Wei-Chen Chiu and Yu-Chiang Frank Wang, Ho coauthors a deep network that summarizes first-person videos, trained on annotated third-person videos plus a small set of annotated first-person ones. The paper goes to arXiv in November 2017 and appears at ECCV 2018.

  2. 2021

    A master's degree at ETH Zurich

    After a BSc in electrical engineering at National Taiwan University, Ho completes a master's degree in computer science at ETH Zurich in 2021. In November 2021, they release Render In-between with Xu Chen, Jie Song and Otmar Hilliges, a BMVC 2021 method that predicts skeletal motion between frames and renders high-frame-rate human video from low-frame-rate footage.

  3. 2023

    Locally editable virtual humans

    As first author with Lixin Xue, Jie Song and Otmar Hilliges, Ho presents Learning Locally Editable Virtual Humans at CVPR 2023, a generative model of neural avatars that stores geometry and texture features on the vertices of a body model so local parts can be swapped between avatars. The team also contributes CustomHumans, a new high-quality dataset for avatar creation and editing.

  4. 2024

    SiTH and 4D-DRESS

    SiTH, a CVPR 2024 pipeline first-authored by Ho with Jie Song and Otmar Hilliges, uses an image-conditioned diffusion model to hallucinate a person's unseen back view and then recovers a full-body textured 3D mesh from the input image and the generated view. With Wenbo Wang as joint first author, they also introduce 4D-DRESS, a real-world 4D clothing dataset of 64 outfits in 520 motion sequences, accepted as a CVPR 2024 highlight.

  5. 2024–2026

    A Taiwan study-abroad scholarship

    From 2024 to 2026, Ho holds the Taiwan Government Scholarship for Studying Abroad, worth USD 32,000. In spring 2025, they serve as a teaching assistant for Machine Perception at ETH Zurich, a course on deep-learning methods for human-centric signals.

  6. 2025

    Personalized body fitting with PHD

    At ICCV 2025, Ho presents PHD as first author, a method that calibrates a user's body shape and then fits 3D poses from video with a shape-conditioned Point Diffusion Transformer prior trained only on synthetic data. With Onat Vuran, a master's thesis student in 2024, they publish ReMu at BMVC 2025, which reconstructs multi-layer 3D clothed humans from images.

  7. 2026

    A doctorate on digital humans

    In February 2026, Ho defends the doctoral thesis "Advancing Digital Humans Toward Personalized Human-Centric AI" in the Advanced Interactive Technologies lab of ETH Zurich's Institute for Intelligent Interactive Systems, supervised by Otmar Hilliges and Marc Pollefeys. Siyu Tang, Ming-Hsuan Yang and Lingjie Liu join Pollefeys on the doctoral committee. In March 2026, they post Gaussian Wardrobe with Zhiyi Chen and colleagues, a 3DV 2026 framework that builds 3D Gaussian avatars from layers of free-form garments for virtual try-on.

  8. As of October 2026

    Research scientist at NVIDIA

    Ho is a research scientist at NVIDIA Research and a member of NVIDIA Research Taiwan, the APAC team of NVIDIA Research, founded in 2022 and led by Frank Wang, which focuses on generative AI, multimodal learning and 3D vision. Their research centers on human-centric computer vision and machine perception powered by generative AI. Earlier, they intern as a research scientist at Meta Reality Labs in Redmond and Google XR in Zurich, working on diffusion models for human pose estimation and 3D human generation.

  1. Google had the most people shaping AI every year from 2014 to 2025Our estimates of which organizations employed the most people shaping AI from 1961 to 2025, with a year-by-year chart.
  2. The United States employs an estimated 53% of people in our data with a known employment countryOur estimates of where AI careers are concentrated now and how countries’ shares of dated positions changed over time.
  3. 46% of living people in our data with known birth and employment countries work outside their birth countryOur estimate of cross-border careers in AI, from country of birth and the country of each current organization.
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionOur estimates of where the people shaping AI studied and where their careers led, from education records in cited biographies.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Hsuan-I Ho yet. Send us a free-licensed photo.

  1. [1711.08922] Summarizing First-Person Videos from Third Persons' Points of Views arxiv.org. Checked 19 October 2026.
  2. Hsuan-I Ho | NVIDIA Research Taiwan research.nvidia.com. Checked 19 October 2026.
  3. Dr. Hsuan-I Ho | AIT Lab ait.ethz.ch. Checked 19 October 2026.
  4. [2111.01029] Render In-between: Motion Guided Video Synthesis for Action Interpolation arxiv.org. Checked 19 October 2026.
  5. Learning Locally Editable Virtual Humans – CVPR 2023 Open Access Repository openaccess.thecvf.com. Checked 19 October 2026.
  6. [2311.15855] SiTH: Single-view Textured Human Reconstruction with Image-Conditioned Diffusion arxiv.org. Checked 19 October 2026.
  7. 4D-DRESS: A 4D Dataset of Real-World Human Clothing With Semantic Annotations – CVPR 2024 Open Access Repository openaccess.thecvf.com. Checked 19 October 2026.
  8. Machine Perception, Spring 2025 | AIT Lab ait.ethz.ch. Checked 19 October 2026.
  9. [2508.21257] PHD: Personalized 3D Human Body Fitting with Point Diffusion arxiv.org. Checked 19 October 2026.
  10. Hsuan-I (Sean) Ho | Personal Webpage azuxmioy.github.io. Checked 19 October 2026.
  11. [2603.04290] Gaussian Wardrobe: Compositional 3D Gaussian Avatars for Free-Form Virtual Try-On arxiv.org. Checked 19 October 2026.
  12. NVIDIA Research Taiwan research.nvidia.com. Checked 19 October 2026.