Fengyu Yang

My name is Fengyu Yang, a first-year PhD student in Computer Science at Yale University, advised by Prof. Alex Wong. My research interest lies in computer vision and multimodal learning, especially in vision and touch.

Previouly, I was an undergraduate student at University of Michigan, advised by Prof. Andrew Owens. I was also fortunate to work with Prof. Wenzhen Yuan, Prof. Xi Li and Prof. Zhongming Liu during my undergraduate study.

Email  /  Github

We introduce UniTouch, a unified tactile representation for vision-based tactile sensors aligned with multiple modalities. We show we can now use powerful models trained on other modalities (e.g. CLIP, LLM) to conduct tactile sensing tasks zero shot.

We present TaRF, a scene representation that brings vision and touch into a shared 3D space.

We propose a unified approach to various touch-related image generation tasks via diffusion model. We propose a novel task of tactile-driven shading estimation and apply our model to existing tasks of visuo-tactile cross generation, and tactile-driven image stylization.

We first consider detection outputs as valuable features for crowd analysis and propose a plug-and-play module Crowd Hat able to boost various detection-based methods.

A dataset of paired vision-and-touch data collected by humans. We apply it to: 1) restyling an image to match a tactile input, 2) self-supervised representation learning, 3) multimodal video prediction.

We first consider the biased context in continue semantic segmentation (CSS) and propose a context-rectified image-duplet learning scheme and a biased-context-insensitive consistency loss to tackle CSS problem.

We propose a prototypical contrastive learning method using both foreground and background categories to tackle the large intra-class variance in geospatial semantic segmentation.