Cited by Ultra-fast light-field microscopy with event detection

Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional Network DOI

D. Li,

Rui Zhong,

Yungang Yang

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(4), P. 991 - 991

Published: Feb. 7, 2025

Light Field Angular Super-Resolution (LFASR) addresses the issue where (LF) images can not simultaneously achieve both high spatial and angular resolution due to limited of optical sensors. Since Spatial-Angular Correlation (SAC) features are closely related structure LF images, its accurate complete extraction is crucial for quality reconstructed by LFASR method based on Deep Neural Networks (DNNs). In low-angular SAC must be extracted from a number pixels that at great distance each other exhibit strong correlations. However, existing methods DNNs fail extract accurately completely. Due receptive field, regular Convolutional (CNNs) unable capture distant pixels, leading incomplete feature extraction. On hand, large convolution kernels attention mechanisms use an excessive features, resulting in insufficient accuracy features. To solve this problem, we introduce Deformable Network (DCN), which adaptively changes position sampling point using offsets, so as pixels. addition, order make offset DCN more further improve also propose Multi-Maximum-Offsets Fusion (MMOF-DCN). MMOF-DCN reduce exploration range finding desired offset, thereby improving efficiency. Experiment results show our proposed has advantages real-world dataset synthetic dataset. The PSNR value have disparity improved 0.45 dB compared methods.

Language: Английский

Citations

High-Speed 3D Imaging DOI

Zijun Ouyang,

Xuge Zhang,

Yutong Li

et al.

Published: Jan. 1, 2025

Language: Английский

Citations

Dynamic Spectral fluorescence microscopy via Event-based & CMOS image-sensor fusion DOI

Robert D. Baird,

Apratim Majumder, Rajesh Menon

et al.

Optics Express, Journal Year: 2024, Volume and Issue: 33(2), P. 2169 - 2169

Published: Dec. 23, 2024

We present a widefield fluorescence microscope that integrates an event-based image sensor (EBIS) with CMOS (CIS) for ultra-fast microscopy spectral distinction capabilities. The EBIS achieves temporal resolution of ∼10 μ s (∼ 100,000 frames/s), while the CIS provides diffraction-limited spatial resolution. A diffractive optical element encodes information into diffractogram, which is recorded by CIS. diffractogram processed using deep neural network to resolve two beads, whose emission peaks are separated only 7 nm and exhibit 88% overlap. validate our imaging capillary flow fluorescent demonstrating significant advancement in microscopy. This technique holds broad potential elucidating foundational dynamic biological processes.

Language: Английский

Citations

Event-based Single Molecule Localization Microscopy (eventSMLM) for High Spatio-Temporal Super-resolution Imaging DOI

Jigmi Basumatary,

S Aravinth,

Neeraj Pant

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Dec. 30, 2023

Photon emission by single molecules is a random event with well-defined distribution. This calls for event-based detection in single-molecule localization microscopy. The detector has the advantage of providing temporal change photons and characteristics within blinking period (typically, ∼ 30 ms ) molecule. information can be used to better localize user-defined collection time (shorter than average time) detector. events collected over every short interval / (∼ 3 give rise several independent photon distributions ( tPSFs experiment showed that intermittently emit photons. So, capturing shorter entire gives realizations PSFs Specifically, this translates sparse active pixels per frame on chip (image plane). Ideally, multiple tPSF position estimates single-molecules, leading centroids. Fitting these centroid points circle provides an approximate (circle center) geometric precision (determined FWHM Gaussian) Since estimate (position precision) directly driven data (photon pixels) recorded , estimated value purely experimental rather theoretical (Thomson’s formula). Moreover, nature camera substantially reduces noise background low-noise environment. method tested three different test samples (1) Scattered Cy3 dye coverslip, (2) Mitochondrial network cell, (3) Dendra2HA transfected live NIH3T3 cells (Influenza-A model). A super-resolution map constructed analyzed based (temporal number photons). Experimental results show 10 nm which 6 fold standard SMLM. imaging HA clustering cellular environment reveals spatio-temporal PArticle Resolution (PAR) (2.3 l p × τ 14.11 par where 1 = −11 meter . second However, brighter probes (such as Cy3) are capable 3.16 Cluster analysis shows > 81% colocalization SMLM, indicating consistency proposed eventSMLM technique. dynamics (migration, association, dissociation) clusters first 60 minutes. With availability high resolution, we envision emergence new kind microscopy particle resolution sub-10 regime.

Language: Английский

Citations

Model-Based Explainable Deep Learning for Light-Field Microscopy Imaging DOI

Pingfan Song, Herman Verinaz-Jadan, Carmel L. Howe

et al.

IEEE Transactions on Image Processing, Journal Year: 2024, Volume and Issue: 33, P. 3059 - 3074

Published: Jan. 1, 2024

In modern neuroscience, observing the dynamics of large populations neurons is a critical step understanding how networks process information. Light-field microscopy (LFM) has emerged as type scanless, high-speed, three-dimensional (3D) imaging tool, particularly attractive for this purpose. Imaging neuronal activity using LFM calls development novel computational approaches that fully exploit domain knowledge embedded in physics and optics models, well enabling high interpretability transparency. To end, we propose model-based explainable deep learning approach LFM. Different from purely data-driven methods, proposed integrates wave-optics theory, sparse representation non-linear optimization with artificial neural network. particular, architecture network designed following precise signal models. Moreover, network's parameters are learned training dataset strategy layer-wise tailored distillation. Such design allows to take advantage new features. It combines benefit both learning-based thereby contributing superior interpretability, transparency performance. By evaluating on structural functional data obtained scattering mammalian brain tissues, demonstrate capabilities achieve fast, robust 3D localization neuron sources accurate identification.

Language: Английский

Citations

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

Sachin Shah, Matthew A. Chan, Haoming Cai

et al.

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Journal Year: 2024, Volume and Issue: 17, P. 25265 - 25275

Published: June 16, 2024

Language: Английский

Citations

光场表征及其分辨率提升技术：文献综述及最新进展(特邀) DOI

张润南 ZHANG Runnan,

周宁 ZHOU Ning,

周子豪 ZHOU Zihao

et al.

Infrared and Laser Engineering, Journal Year: 2024, Volume and Issue: 53(9), P. 20240347 - 20240347

Published: Jan. 1, 2024

Citations

Ultra-fast light-field microscopy with event detection DOI

Liheng Bian, Xuyang Chang, Hanwen Xu

et al.

Light Science & Applications, Journal Year: 2024, Volume and Issue: 13(1)

Published: Nov. 7, 2024

The event detection technique has been introduced to light-field microscopy, boosting its imaging speed in orders of magnitude with simultaneous axial resolution enhancement scattering medium.

Language: Английский

Citations