Yu-Hsuan Yeh

I got my second M.S. degree at CMU MSCV.

I received my first M.Sc. at National Yang Ming Chiao Tung University(NYCU) advised by Prof. Wei-Chen Chiu and co-advised by Prof. Min Sun and Dr. Yi-Hsuan Tsai.

I am working on computer vision problems related to panoramic images for layout and depth applications.

I am actively searching for full-time positions beginning in 2024.

Please feel free to drop me an email if you're interested in my background!

Email: yuhsuany@andrew.cmu.edu
Resume  /  Google Scholar  /  Github  /  Linkedin

My research interest mainly focuses on deep learning and computer vision. I am currently working on projects related to scene understanding, such as depth and layout estimation. My goal is to build enriched vision applications and reliable AI systems that can perceive various environments and produce robust estimations for humans.

We propose a new fusion module and Contrast-Aware Photometric Loss to improve the performance of BiFuse and increase the stability of self-training on real-world videos.

We propose a differentiable layout-to-depth procedure to convert the 360 layout representation into the 360 horizon-depth map, thus enabling the training objective for our layout estimation network to take advantage of 3D geometric information.

We propose a two-branch neural network leveraging two common projections – equirectangular and cubemap projections – as inputs to predict the depth map of a monocular 360 image.

We release the first real-world dataset containing paired depth and layout annotations.