Article
Optics
Haiquan Hu, Hao Zhou, Zhihai Xu, Qi Li, Huajun Feng, Yueting Chen, Tingting Jiang, Wenbin Xu
Summary: This study proposed a practical hyperspectral camera with a designed diffractive optical element (DOE), which achieved good spatial resolution and spectral accuracy in snapshot spectral imaging. This system had great potential in portable hyperspectral imaging system.
OPTICS AND LASERS IN ENGINEERING
(2022)
Article
Optics
Yingwen Zhang, Duncan England, Benjamin Sussman
Summary: This article presents a snapshot hyperspectral imaging technique that utilizes quantum ghost spectroscopy, allowing high spectral resolution without sacrificing spatial resolution and improving resource efficiency.
Article
Optics
Tingkui Mu, Feng Han, Haoyang Li, Abudusalamu Tuniyazi, Qiuxia Li, Hang Gong, Wenjing Wang, Rongguang Liang
Summary: The snapshot optically replicating and remapping imaging spectropolarimeter (ORRISp) combines spectrally-modulated polarimetry with optical replicating and remapping imaging to recover spectrally-resolved Stokes parameters from a 2D scene. The prototype demonstrates spectral and spectropolarimetric imaging performance, covering a spectral range of 450 - 750 nm with 56 channels and a field of view of +/- 5.8 degrees, operating at video rate.
OPTICS AND LASERS IN ENGINEERING
(2022)
Review
Medicine, General & Internal
Manuel Barberio, Sara Benedicenti, Margherita Pizzicannella, Eric Felli, Toby Collins, Boris Jansen-Winkeln, Jacques Marescaux, Massimo Giuseppe Viola, Michele Diana
Summary: Hyperspectral imaging (HSI) is a novel optical imaging modality that allows for contactless and non-destructive biochemical analysis of living tissue at a molecular level. It has diverse applications in the medical field and provides quantitative and qualitative information to objectively discriminate between different tissue types and between healthy and pathological tissue. Despite its potential, HSI is still not widely used in daily surgical practice, but has recently been utilized as an intraoperative guidance tool within different surgical disciplines.
Article
Optics
Nan Xu, Hao Xu, Shiqi Chen, Haiquan Hu, Zhihai Xu, Huajun Feng, Qi Li, Tingting Jiang, Yueting Chen
Summary: Hyperspectral imaging attempts to improve coding aperture design in phase-coded systems and introduces equalization designed phase-coded apertures using wave optics for richer features in image reconstruction. Our CAFormer network achieves better image reconstruction results with less computation by substituting self-attention with channel-attention. This work optimizes the imaging process from hardware design, reconstruction algorithm, and PSF calibration, bringing snapshot compact hyperspectral technology closer to practical applications.
Article
Computer Science, Artificial Intelligence
Lishun Wang, Miao Cao, Yong Zhong, Xin Yuan
Summary: Video snapshot compressive imaging (SCI) is a computational imaging technique that captures multiple video frames with a single measurement. It utilizes masks to modulate high-speed frames, which are then summed to a single measurement captured by a low-speed 2D sensor. This article focuses on the reconstruction algorithm for video SCI and introduces a Spatial-Temporal transFormer (STFormer) that exploits correlation in spatial and temporal domains. The STFormer network, consisting of token generation block, video reconstruction block, and multiple STFormer blocks, achieves state-of-the-art performance in both simulated and real data.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Engineering, Electrical & Electronic
Yubao Sun, Xunhao Chen, Mohan S. Kankanhalli, Qingshan Liu, Junxia Li
Summary: This paper proposes a residual ensemble network to improve the reconstruction quality in video snapshot compressive imaging (SCI) system by learning the spatiotemporal correlations. The network effectively fuses the predictions of sub-networks and maintains the consistency between video frames. Experimental results show that the network significantly improves the reconstruction quality while maintaining low computational cost.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Materials Science, Multidisciplinary
Zhihan Hong, Yuanyuan Sun, Piaoran Ye, Douglas A. Loy, Rongguang Liang
Summary: A 3D printed glass lightguide array is developed to address the challenges of high spatial resolution in snapshot hyperspectral imaging. It samples the intermediate image in high spatial resolution and redistributes the pixels to achieve high spectral resolution. This technology simplifies the imaging system, reduces complexity and cost, and has demonstrated good performance with biological samples. It will catalyze the development of new hyperspectral imaging systems and enable new applications from UV to infrared.
ADVANCED OPTICAL MATERIALS
(2023)
Article
Computer Science, Artificial Intelligence
Xin Yuan, Yang Liu, Jinli Suo, Fredo Durand, Qionghai Dai
Summary: This article investigates the reconstruction problem in video snapshot compressive imaging (SCI) and proposes fast and flexible algorithms based on the plug-and-play framework. By applying image and video deep denoising priors, the algorithms successfully recover high-definition color videos and are extended to color SCI systems. Results from simulation and real datasets validate the superiority of the proposed algorithm.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Optics
Hui Xie, Zhuang Zhao, Jing Han, Yi Zhang, Lianfa Bai, Jun Lu
Summary: In this paper, a method is proposed that uses optical imaging and convolutional neural networks to solve the snapshot hyperspectral imaging reconstruction problem. Unlike traditional supervised deep learning methods, this method does not require training and performs well in different real-life scenarios.
OPTICS AND LASERS IN ENGINEERING
(2022)
Article
Chemistry, Analytical
Cory Juntunen, Isabel M. Woller, Yongjin Sung
Summary: Hyperspectral 3D imaging integrates SPOT and FTS technologies for acquiring projection images from different viewing angles instantly and performing high-spectral-resolution imaging, which is validated through imaging performance using fluorescent beads and sunflower pollens.
Article
Gastroenterology & Hepatology
Robert Sucher, Uwe Scheuermann, Sebastian Rademacher, Andri Lederer, Elisabeth Sucher, Hans-Michael Hau, Gerald Brandacher, Stefan Schneeberger, Ines Gockel, Daniel Seehofer
Summary: This study evaluated the utility of hyperspectral imaging (HSI) in monitoring the microcirculation of the graft and adequate perfusion of the intestinal anastomosis during pancreatic allotransplantation. The results showed that HSI was effective in monitoring graft perfusion and oxygenation, potentially improving the outcome of pancreas transplantation.
HEPATOBILIARY SURGERY AND NUTRITION
(2022)
Article
Computer Science, Artificial Intelligence
Zongliang Wu, Chengshuai Yang, Xiongfei Su, Xin Yuan
Summary: In this paper, an online deep optimization algorithm is proposed for video snapshot compressive imaging, which has better adaptability and performance. Furthermore, a deep demosaicing prior is introduced to address the challenge of color video imaging.
INTERNATIONAL JOURNAL OF COMPUTER VISION
(2023)
Article
Engineering, Electrical & Electronic
Yubao Sun, Junru Huang, Liling Zhao, Kai Hu
Summary: The hyperspectral snapshot compressive imaging system utilizes compressive sensing to reconstruct three-dimensional hyperspectral images. The proposed encoder and decoder network with dense back-projection joint attention and spatial-spectral attention modules significantly improve reconstruction efficiency and quality.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Ying Yang, Yong Xie, Xunhao Chen, Yubao Sun
Summary: Snapshot Compressive Imaging is a technology based on compressive sensing theory for high-efficiency hyperspectral data acquisition. This paper proposes a novel deep network for reconstructing 3D hyperspectral data from 2D snapshot measurements, utilizing symmetric residual and non-local spatial-spectral attention modules. Experimental results demonstrate the proposed network outperforms competing methods in terms of reconstruction quality and running time.