I-Ting Tsai

I-Ting Tsai

Hi, I'm a 4th-year CS PhD student at Cornell University, advised by Prof. Bharath Hariharan. I work on computer vision and generative AI, with a focus on 3D and video generation — bridging my background in architecture/design with deep learning.

Contact: itingtsai@cs.cornell.edu

LinkedIn GitHub Google Scholar

Research

3D Synthesis for Architectural Design

3D Synthesis for Architectural Design

I-Ting Tsai, Bharath Hariharan

WACV, 2025 (Application Track)

TL;DR

We build the first training-free 3D synthesis pipeline for architectural design — generating geometry, texture, and detail separately to produce diverse, clean, easily-editable buildings, preferred 95% of the time for diversity and 84% for realism over the strongest baseline.

Short Papers

Million Eyes on the Robot Ump

Million Eyes on the "Robot Ump": The Case for Studying Sports in HRI Through Baseball

Waki Kamino, Andrea Wang, Dhruv Agarwal, Sil Hamilton, Eun Jeong Kang, Jieun Kim, Keigo Kusumegi, Pegah Moradi, Daniel Mwesigwa, Yan Tao, I-Ting Tsai, Ethan Yang, Shengqi Zhu, Shu-Jung Han, Chi-Jung Lee, Michael Joseph Sack, Tianhong Catherine Yu, Weslie Khoo, Andy Elliot Ricci, Yoyo Tsung-Yu Hou, Boyoung Kim, Dr. Selma Sabanovic, David Crandall, Karen Levy, Malte F Jung

HRI, 2025

TL;DR

A position paper making the case for sports as a real-world testbed for human-robot interaction, using MLB's vision-based "robot umpire" to surface emerging questions in power dynamics, labor, and the technical challenges of large-scale computer vision deployment.


Events

2026 Jun Workshop Synthetic Data for Computer Vision @ CVPR 2026