By leveraging classical coded exposure imaging technique and emerging implicit neural representation for videos, we develop a novel self-recursive neural network to sequentially retrieve the latent video sequence from the blurry image utilizing the embedded motion direction cues.
@article{zhangBDINR2024,title={Lightweight {{High-Speed Photography Built}} on {{Coded Exposure}} and {{Implicit Neural Representation}} of {{Videos}}},author={Zhang, Zhihong and Yang, Runzhao and Suo*, Jinli and Cheng, Yuxiao and Dai, Qionghai},year={2024},journal={International Journal of Computer Vision},issn={0920-5691, 1573-1405},doi={10.1007/s11263-024-02198-1},}
This paper provides a review of the advancements in video compressive sensing over the past decade. Research gaps and future directions towards real-world applications are put forward as well.
@article{ZHANG2024SCI,title={A Decade Review of Video Compressive Sensing: A Roadmap to Practical Applications},journal={Engineering},author={Zhangᐪ, Zhihong and Zhengᐪ, Siming and Qiu, Min and Situ, Guohai and Brady, David J. and Dai, Qionghai and Suo*, Jinli and Yuan*, Xin},year={2024},issn={2095-8099},doi={https://doi.org/10.1016/j.eng.2024.08.013},}
We proposed an end-to-end framework to handle general motion blurs with a unified deep neural network, and optimize the shutter’s encoding pattern together with the deblurring processing to achieve high-quality sharp images.
title = {Deep Coded Exposure: End-to-End Co-optimization of Flutter Shutter and Deblurring Processing for General Motion Blur Removal},author = {Zhang, Zhihong and Dong, Kaiming and Suo*, Jinli and Dai, Qionghai},journal = {Photonics Research},year = {2023},volume = {11},number = {10},pages = {1678},doi = {10.1364/PRJ.489989},}
Under low-light environment, handheld photography suffers from severe camera shake under long exposure settings. Sophisticated noise and saturation regions are two dominating challenges in practical low-light deblurring. In this work, we propose a novel non-blind deblurring method dubbed image and feature space Wiener deconvolution network (INFWIDE) to tackle these problems systematically.
@article{infwide_zhihong,title={{{INFWIDE}}: {{Image}} and Feature Space Wiener Deconvolution Network for Non-blind Image Deblurring in Low-Light Conditions},author={Zhang, Zhihong and Cheng, Yuxiao and Suo*, Jinli and Bian, Liheng and Dai, Qionghai},journal={IEEE Transactions on Image Processing},year={2023},volume={32},pages={1390--1402},doi={10.1109/TIP.2023.3244417},}
In this paper, we propose to build a dual-sensor camera to additionally collect the photons in NIR wavelength, and make use of the correlation between RGB and near-infrared (NIR) spectrum to perform high-quality reconstruction from noisy dark video pairs.
@article{CHENG2023429,title={A mutually boosting dual sensor computational camera for high quality dark videography},journal={Information Fusion},volume={93},pages={429-440},year={2023},issn={1566-2535},doi={https://doi.org/10.1016/j.inffus.2023.01.013},author={Cheng, Yuxiao and Yang, Runzhao and Zhang, Zhihong and Suo*, Jinli and Dai*, Qionghai},keywords={Computational photography, Video denoising, Low light video, Dark vision, RGB-NIR, Dual-channel network}}
A novel framework that combines computational imaging (CI) and computer vision (CV) to retrieve semantic vison information with low bandwidth.
@article{zhang2022CompressiveSampling,title={From Compressive Sampling to Compressive Tasking: Retrieving Semantics in Compressed Domain with Low Bandwidth},shorttitle={From Compressive Sampling to Compressive Tasking},author={Zhangᐪ, Zhihong and Zhangᐪ, Bo and Yuanᐪ, Xin and Zheng, Siming and Su, Xiongfei and Suo*, Jinli and Brady, David J. and Dai*, Qionghai},year={2022},journal={PhotoniX},volume={3},number={1},pages={19},}
A deep optics based snapshot compressive imaging (SCI) system for simultaneous spatial-temporal super-resolution.
@article{zhang2022EndtoendSnapshot,title={End-to-End Snapshot Compressed Super-Resolution Imaging with Deep Optics},author={Zhangᐪ, Bo and Yuanᐪ, Xin and Deng, Chao and Zhang, Zhihong and Suo*, Jinli and Dai*, Qionghai},year={2022},journal={Optica},volume={9},number={4},pages={451--454},publisher={{Optica Publishing Group}},}
Unsupervised LSTM/CNN based event sequence denoisng algorithms for dynamic vision sensors.
@inproceedings{Event_denoise,title={Denoising of Event-Based Sensors with Deep Neural Networks},booktitle={Optoelectron. {{Imaging Multimed}}. {{Technol}}. {{VIII}}},author={Zhang, Zhihong and Suo*, Jinli and Dai, Qionghai},year={2021},volume={11897},pages={203--209},publisher={{SPIE}},organization={{International Society for Optics and Photonics}},}
A multi-channel deep U-Net for high axial resolution single molecule localization (SML) under dense excitation.
@article{zhang2021HighaxialresolutionSinglemolecule,title={High-Axial-Resolution Single-Molecule Localization under Dense Excitation with a Multi-Channel Deep {{U-Net}}},author={Zhang, Weihang and Zhang, Zhihong and Bian, Liheng and Wang, Haoqian and Suo*, Jinli and Dai, Qionghai},year={2021},journal={Optics Letters},volume={46},number={21},pages={5477--5480},publisher={{Optical Society of America}},}
A 10-mega-pixel high-throughput high-speed snapshot compressive imaging (SCI) system with a hybrid coded aperture.
@article{zhang2021TenmegapixelSnapshot,title={Ten-Mega-Pixel Snapshot Compressive Imaging with a Hybrid Coded Aperture},author={Zhangᐪ, Zhihong and Dengᐪ, Chao and Liu, Yang and Yuan, Xin and Suo*, Jinli and Dai, Qionghai},year={2021},journal={Photonics Research},volume={9},number={11},pages={2277--2287},}