Haoming Cai /haʊˈmɪŋ tsaɪ/

I am a second-year CS PhD student @ CS Department of University of Maryland, College Park, advised by Professor Christopher Metzler.

Previously, I had a wonderful journey (2020-2022) in X-Pixel with my supervisor Prof. Dong Chao (Shanghai AI Lab) and mentor Dr. Gu Jinjin (Univeristy of Sydney).

Contact : hmcai@umd.edu / helmut.choy@gmail.com / haomingcai@link.cuhk.edu.cn

My CV : PDF

GitHub  /  Google Scholar  / 

profile photo

Research Interest

- My research centers on the convergence of Computational Photography, Gen AI and Low-level Vision. I aim to synergize computational imaging techniques with advanced back-end processing algorithms to enhance the perceptual quality of human experiences on mobile and edge devices.
Sometimes Science Is More Art Than Science -- Rick Sanchez



3D Reconstruction in Computational Photography (2023-Now)

project image

Flash-Splat: 3D Reflection Removal with Flash Cues and Gaussian Splats


Mingyang Xie*, Haoming Cai* , Sachin Shah, Yiran Xu, Brandon Y. Feng, Jia-bin Huang, Christopher Metzler
Under Review - 2024




Seeing & Tracking Through Adverse Weather Condition (2023-Now)

project image

Temporally Consistent Atmospheric Turbulence Mitigation with Neural Representations


Haoming Cai*, Jingxi Chen*, Brandon Y. Feng, Weiyun Jiang, Mingyang Xie, Kevin Zhang, Cornelia Fermuller, Yiannis Aloimonos, Ashok Veeraraghavan, Christopher Metzler.
Under Review - 2024

project image

CodedEvents: Optimal Point-Spread-Function Engineering for 3D-Tracking with Event Cameras


Sachin Shah, Matthew Albert Chan, Haoming Cai, Jingxi Chen, Sakshum Kulshrestha, Chahat Deep Singh, Yiannis Aloimonos, Christopher Metzler.
The IEEE / CVF Computer Vision and Pattern Recognition Conference (CVPR'24)

This paper explores PSF engineering for neuromorphic event cameras, designing optimized masks for superior 3D point localization and tracking.

project image

Snow Removal in Video: A New Dataset and A Novel Method


Haoyu Chen, Jingjing Ren, Jinjin Gu, Hongtao Wu, Xuequan Lu, Haoming Cai, Lei Zhu
Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV'23)
arxiv/ code / website /

This paper presents a new deep learning framework for removing snow from videos, featuring a high-quality dataset and innovative modules for effective snow removal, outperforming existing methods.




Image/Video Generation through Diffusion Model (2023-Now)

project image

TimeRewind: Rewinding Time with Image-and-Events Video Diffusion


Jingxi Chen, Brandon Y. Feng, Haoming Cai, Mingyang Xie, Christopher Metzler, Cornelia Fermuller, Yiannis Aloimonos
Under Review - 2024

Image Metric Design Inspired by Human Visual System (2020-Now)

project image

Assessor360: Multi-sequence Network for Blind Omnidirectional Image Quality Assessment


Tianhe Wu, Shuwei Shi, Haoming Cai , Mingdeng Cao, Jing Xiao, Yinqiang Zheng, Yujiu Yang
Advances in Neural Information Processing Systems (NeurIPS'23)
arxiv/ code / website /

Current omnidirectional image quality assessment lacks observer browsing modeling. We propose Assessor360, a novel multi-sequence network for BOIQA derived from realistic multi-assessor ODI quality assessment

project image

Pipal: a large-scale image quality assessment dataset for perceptual image restoration


Jinjin Gu, Haoming Cai, Haoyu Chen, Xiaoxing Ye, Jimmy S Ren, Chao Dong.
The European Conference on Computer Vision (ECCV'20)
arxiv/ website /

This paper highlights the challenge IQA faces with emerging GAN-based image restoration methods, noting a growing disparity between quantitative metrics and perceptual quality. To address this, the authors introduce a large-scale IQA dataset and benchmarks to enhance IQA methods’ effectiveness.

Controllable & Efficient Image Restoration (2020 - 2022)

project image

Super-resolution by predicting offsets: An ultra-efficient super-resolution network for rasterized images


Jinjin Gu, Haoming Cai, Chenyu Dong, Ruofan Zhang, Yulun Zhang, Wenming Yang, Chun Yuan.
European Conference on Computer Vision (ECCV'22)
arxiv/ code /

SRPO is a real-time super-resolution method for computer graphics, achieving superior visual effects with minimal computational cost by leveraging rasterized image features and offset prediction.

project image

Efficient image super-resolution using vast-receptive-field attention


Haoming Cai*, Lin Zhou*, Jinjin Gu, Zheyuan Li, Yingqi Liu, Xiangyu Chen, Yu Qiao, Chao .
AIM'22 @ ECCV
arxiv/ code /

This study improves super-resolution networks by refining the attention mechanism, leading to VapSR, which outperforms lightweight networks with fewer parameters, achieving similar results as IMDB and RFDN networks with significantly fewer parameters.

project image

Blueprint separable residual network for efficient image super-resolution


Zheyuan Li, Yingqi Liu, Xiangyu Chen, Haoming Cai, Jinjin Gu, Yu Qiao, Chao Dong.
NTIRE'23 @ CVPR
arxiv/ code /

Winner of Efficient Image Super-Resolution Track @ New Trends in Image Restoration and Enhancement (NTIRE) workshop of CVPR’23

project image

Toward interactive modulation for photo-realistic image restoration


Haoming Cai, Jingwen He, Yu Qiao, Chao Dong.
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshop (CVPRW'21)
arxiv/ code /

This paper presents CUGAN, a Controllable Unet GAN, for modulating image restoration tasks with fine texture details. Through dynamic level adjustments and condition networks, CUGAN outperforms previous methods, offering smooth user control over output effects.


Design and source code from Jon Barron's website