STEC: A Spatio-Temporal Entropy Coverage Metric for Evaluating Sampled Video Frames
Shih-Yao Lin
WACV Workshop, 2026
From Captions to Keyframes: KeyScore for Multimodal Frame Scoring and Video-Language Understanding
Shih-Yao Lin, Sibendu Paul, Caren Chen
arXiv preprint, arXiv: 2510.06509
Cross-Lingual Profanity Detection Using an Artificial Intelligence Framework
Shih-Yao Lin and Kewen Chen
U.S. Patent Application, submitted 2025 (Under Review)
Modality-Specific and Modular Framework for Multimodal Artificial Intelligence
Shih-Yao Lin and Kewen Chen
U.S. Patent Application, submitted 2025 (Under Review)
Training Architecture and Techniques with Context Using Multimodal Input
Shih-Yao Lin, Minmin Shen, Kewen Chen
U.S. Patent Application, submitted 2025 (Under Review)
TriMoE-VLM: A Vision-Language MoE Framework for Tri-Modal Contrastive Learning with Video, Text, and Fused Multimodal Embeddings
Shih-Yao Lin, Minmin Shen, Kewen Chen
U.S. Patent Application, submitted 2025 (Under Review)
Context-Aware Video Reframing Using Multimodal Scene Understanding
Shih-Yao Lin, Sibendu Paul, Kewen Chen
U.S. Patent Application, submitted 2025 (Under Review)
Systems and Methods for Context-Aware Video Reframing Using Multimodal SceneUnderstanding
Minmin Shen, ..., Shih-Yao Lin, ..., Kewen Chen
U.S. Patent Application, submitted 2025 (Under Review)
Video-based 3D Hand Pose and Mesh Estimation Based on Temporal-Aware Self-Supervised Learning
Shih-Yao Lin, Yusheng Xie, Hui Tang, Chao Huang, Lianyi Han, and Wei Fan
US Patent (pending)
Synthesizing 3D Hand Pose Based On Multi-modal Guided Generative Networks
Shih-Yao Lin, Yusheng Xie, Hui Tang, Chao Huang, Lianyi Han, and Wei Fan
U.S. Patent No. 11610326 B2, 2023
Shih-Yao Lin, Yusheng Xie, Hui Tang, Lianyi Han,and Wei Fan
US Patent, US10916050B1, 2021 [PDF]
Vision-based rehabilitation training system based on 3d human pose estimation using multi-view images
Shih-Yao Lin, Tao Yang, Chao Huang, Zhen Qian, Wei Fan
U.S. Patent Application No. 20220148453 A1, Published 2022
Augmenting reliable training data with CycleGAN for hand pose estimation
Shih-Yao Lin, Yusheng Xie, Kun Wang, Lianyi Han, and Wei Fan
US Patent, US20200372668A1, 2020 [PDF]