Zhixiang Wang | 汪智祥

I am a MEXT-funded PhD student at the University of Tokyo with Shin'ichi Satoh starting in Oct. 2021. Our lab is affiliated with National Institute of Informatics. Part of my research was supported by RIISE. Prior to this, I completed my Master's degree at National Taiwan University with Yung-Yu Chuang.

My research interest focuses on computational imaging and computational photography. Throughout my studies, I have been fortunate to collaborate closely with Yinqiang Zheng, Jia-Bin Huang, Jian Wang, and Jinwei Gu. In 2022, I interned with Shree Nayar's team at Snap Research. During my master's study, I was fortunate to work at National Institute of Informatics as a research intern. I also spent some time at Nanyang Technological University in Singapore.

The
News
Research

My research is centered on Computational Photography and Imaging, with a focus on capturing the intricacies of the physical world through innovative hardware and algorithmic designs. I am now keen on blending physical principles with generative models for reconstructions and interactions of the physical world.

DiffIR2VR-Zero: Zero-Shot Video Restoration with Diffusion-based Image Restoration Models

Chang-Han Yeh Chin-Yang Lin Zhixiang Wang Chi-Wei Hsiao Ting-Hsuan Chen Yu-Lun Liu

arXiv 2024

This paper introduces a method for zero-shot video restoration using pre-trained image restoration diffusion models.

Matting by Generation

Zhixiang Wang Baiang Li Jian Wang Yu-Lun Liu Jinwei Gu Yung-Yu Chuang Shin'ichi Satoh

The ACM Special Interest Group on Computer Graphics (SIGGRAPH) 2024

We repurpose the pre-trained text-to-image diffusion model for image matting.

DisCO: Portrait Distortion Correction with Perspective-Aware 3D GANs

Zhixiang Wang Yu-Lun Liu Jia-Bin Huang Shin'ichi Satoh Sizhuo Ma Guru Krishnan Jian Wang

International Journal of Computer Vision (IJCV) 2024

We build a pipeline to use pre-trained 3D GANs for correcting face perspective distortion in close-up portraits.

HumanNeRF-SE: A Simple yet Effective Approach to Animate HumanNeRF with Diverse Poses

Caoyuan Ma Yu-Lun Liu Zhixiang Wang Wu Liu Xinchen Liu Zheng Wang

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2024

We propose a simple yet effective HumanNeRF for animating human with diverse poses.

Rethinking Video Frame Interpolation from Shutter Mode Induced Degradation

Xiang Ji Zhixiang Wang Zhihang Zhong Yinqiang Zheng

International Conference on Computer Vision (ICCV) 2023

We propose an unified method for motion from temporal degradations caused by different shutter modes.

HOTCOLD Block: Fooling Thermal Infrared Detectors with a Novel Wearable Design

Hui Wei* Zhixiang Wang* Xuemei Jia Yinqiang Zheng Hao Tang Shin'ichi Satoh Zheng Wang

AAAI Conference on Artificial Intelligence (AAAI) 2023

We use anti-fever stickers/heating pads to form an effective solution to protect human privacy under intelligent thermal cameras.

Single Image Deblurring with Row-dependent Blur Magnitude

Xiang Ji Zhixiang Wang Shin'ichi Satoh Yinqiang Zheng

International Conference on Computer Vision (ICCV) 2023

We propose a method for removing row-wise motion blur with single image input.

Neural Global Shutter: Learn to Restore Video from a Rolling Shutter Camera with Global Reset Feature

Zhixiang Wang Xiang Ji Jia-Bin Huang Shin'ichi Satoh Xiao Zhou Yinqiang Zheng

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2022

We offer an alternative and new solution for RS rectification by turning the rectification problem into a deblur-like one with the widely ignored RSGR feature.

Both Style and Fog Matter: Cumulative Domain Adaptation for Semantic Foggy Scene Understanding

Xianzheng Ma Zhixiang Wang Yacheng Zhan Yinqiang Zheng Zheng Wang Dengxin Dai Chia-Wen Lin

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2022, Oral

We alleviate the domain gap caused by mixed fog influence and style variation without labels.

Domain-Specific Mappings for Generative Adversarial Style Transfer

Hsin-Yu Chang Zhixiang Wang Yung-Yu Chuang

European Conference on Computer Vision (ECCV) 2020

We propose domain-specific mappings for remapping latent features in the shared content space to domain-specific content spaces.

Learning to Reduce Dual-level Discrepancy for Infrared-Visible Person Re-identification

Zhixiang Wang Zheng Wang Yinqiang Zheng Yung-Yu Chuang Shin'ichi Satoh

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2019

We propose a dual-level discrepancy reduction learning scheme to separate the modality and appearance.

Polarimetric Camera Calibration Using an LCD Monitor

Zhixiang Wang Yinqiang Zheng Yung-Yu Chuang

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2019

We jointly calibrate the polarizer angles and the inverse CRF using a slightly adapted checker pattern displayed on an LCD monitor.