This paper presents findings from a spaceborne Earth observation experiment utilizing novel, ultra-compact hyperspectral imaging camera aboard 3U CubeSat. Leveraging the Offner optical scheme, camera’s hyperspectrometer captures images of terrestrial regions with 200 m spatial resolution and 12 nanometer spectral across 400 to 1000 wavelength range, covering 150 channels in visible near-infrared spectrums. The is specifically designed for deployment on CubeSat nanosatellite platform,...

Neurofeedback based on real-time functional magnetic resonance imaging (rt-fMRI) is a novel and rapidly developing research field. It allows for training of voluntary control over localized brain activity connectivity has demonstrated promising clinical applications. Because the rapid technical developments MRI techniques availability high-performance computing, new methodological advances in rt-fMRI neurofeedback become possible. Here we outline core components open-source framework, termed...

In recent years, several pioneering works were dedicated to imaging systems based on simple diffractive structures like Fresnel lenses or phase zone plates. Such are much lighter and cheaper than classical refractive optical systems. However, the quality of images obtained by optics suffers from stronger distortions various types. this paper, we show that a combination high-precision lens design with post-capture computational reconstruction allows one attain higher image quality. The...

Launched in March 2021, the 3U CubeSat nanosatellite was first ever to use an ultra-lightweight harmonic diffractive lens for Earth remote sensing. We describe platform we used; our 10 mm diameter and 70 focal length synthesis, design, manufacturing; a custom 3D-printed camera housing built from zero-thermal-expansion metal alloy; on-Earth image post-processing with convolutional neural network resulting images comparable quality classical refractive optics used sensing before.

In this paper, we describe our advances in manufacturing a 256-layer 7-μm thick harmonic lens with 150 and 300 mm focal distances combined color correction, deconvolution, feedforwarding deep learning neural network capable of producing images approaching photographic visual quality. While reconstruction taken diffractive optics was presented previous works, paper is the first to use networks during restoration step. The level imaging quality achieved system can facilitate emergence...

Artificial intelligence (AI) significantly enhances the development of Meta-Optics (MOs), which encompasses advanced optical components like metalenses and metasurfaces designed to manipulate light at nanoscale. The intricate design these requires sophisticated modeling optimization achieve precise control over behavior, tasks for AI is exceptionally well-suited. Machine learning (ML) algorithms can analyze extensive datasets simulate numerous variations identify most effective...

The synergy between artificial intelligence (AI) and hyperspectral imaging (HSI) holds tremendous potential across a wide array of fields. By leveraging AI, the processing interpretation vast complex data generated by HSI are significantly enhanced, allowing for more accurate, efficient, insightful analysis. This powerful combination has to revolutionize key areas such as agriculture, environmental monitoring, medical diagnostics providing precise, real-time insights that were previously...

Many recent studies have focused on developing image reconstruction algorithms in optical systems based flat optics. These demonstrate the feasibility of applying a combination optics and real vision systems. However, additional causes quality loss been encountered development such This study investigates influence reconstructed factors as limitations mass production technology for diffractive optics, lossy video stream compression artifacts, specificities neural network approach to...

We propose a method of statistical shape modeling applied to reconstructing anatomical structures with deformations. This is promising for deformed body parts that have certain normal shape. The allows one reconstruct object using information about the part and its undeformed fragment, while taking into account peculiar individual features variability relative average

The article considers a new type of equipment for unmanned aerial vehicles based on the use diffractive optical elements imaging and deep learning subsequent real-time image quality improvement.

In this paper, we propose an approach to the classification of high-resolution hyperspectral images in applied problem identification vegetation types. A modified spectral-spatial convolutional neural network with compensation for illumination variations is used as a classifier. For generating training dataset, algorithm based on adaptive index proposed. The effectiveness proposed shown basis survey data agricultural lands obtained from compact camera developed in-house.

In this paper, we present a hybrid refractive-diffractive lens that, when paired with deep neural network-based image reconstruction, produces high-quality, real-world images minimal artifacts, reaching PSNR of 28 dB on the test set. Our diffractive element compensates for off-axis aberrations single refractive and has reduced chromatic across visible light spectrum. We also describe our training set augmentation novel quality criteria called “false edge level” (FEL), which validates that...

The paper presents a study of various approaches to the classification soil covers based on neural network algorithms using hyperspectral remote and proximal sensing Earth. spectral distributions were recorded in laboratory an Offner imaging scanning hyperspectrometer. Spectral-spatial characteristics nine samples from parts farming land Samara region experimentally studied. Using method energy dispersion microanalysis, correspondence between data chemical composition taken was established....

Quality assessment and artifact detection in functional magnetic resonance imaging (fMRI) data is essential for clinical applications brain research. Subject head motion remains the main source of artifacts - even tiniest movement can perturb structural derived from fMRI. In this paper, we propose an end-to-end neural network technology detecting step anomalies with training on partially synthetic adaptation to a specific small set real data. A procedure generating dataset module automated...

In traditional neural network designs, a multilayer perceptron (MLP) is typically employed as classification block following the feature extraction stage. However, Kolmogorov-Arnold Network (KAN) presents promising alternative to MLP, offering potential enhance prediction accuracy. this paper, we studied KAN-based networks for pixel-wise of hyperspectral images. Initially, compared baseline MLP and KAN with varying numbers neurons in their hidden layers. Subsequently, replaced linear,...

This paper describes a unified approach to correct optical distortions in images formed by Fresnel lens with computational post-processing that opens up new opportunities use lenses lightweight and inexpensive computer vision devices. Traditional methods of aberration correction do not address artifacts introduced systematic way thus fail deliver image quality acceptable for general-purpose color imaging. In our approach, the is restored using three steps: first, deblurring base channel,...

The possibility of essentially reducing the weight and production cost computer vision systems has led to publication a large number research works dealing with development new imaging based on diffractive optics. This study proposes system composed three lenses, each forming separate channel color RGB image. approach allows us significantly narrow spectral range lens, thus image distortion caused by chromatic aberration inherent in It shows that this scheme perform neural network-aided...

A hyperspectrometer based on the Offner scheme was investigated. Spectral characteristics were studied and calibrated using a standard spectrometer. As result of estimating deviations spectra imaging reference spectrometer, calibration coefficients obtained. The reflectance beets, onions potatoes under natural solar illumination experimentally Based analysis hyperspectral data, an distribution vegetative indices and, in particular, moisture content, carried out. Analysis histograms content index

With suggested computational post-processing workflow for correcting optical distortions, the Fresnel lens can finally be used in lightweight and inexpensive computer vision sensors. Common methods image enhancement do not comprehensively address blurring artifacts caused by strong chromatic aberrations images produced a simple system. To deliver quality acceptable general-purpose color imaging, we propose post-capture processing to enhance of acquired with 256-level lens. The PSNR measure...

The pressure to reduce weight and improve image quality of the imaging devices continues push research in area flat optics with computational reconstruction. This paper presents a new end-to-end framework applying two convolutional neural networks (CNNs) reconstruct images captured multilevel diffractive lenses (MDLs). We show that patch-wise chromatic blur image-wise context-aware color highlights, distortions inherent MDLs, can be successfully addressed suggested reconstruction pipeline....

This paper examines the effectiveness of neural network algorithms for hydraulic system fault detection and a novel architecture is suggested. The proposed gated convolutional autoencoder was trained on simulated training set augmented with just 0.2% data from real test bench, dramatically reducing time needed to spend actual hardware build high-quality model. Our model validated bench showed accuracy more than 99% correctly recognized states 10-s sampling window. can be also leveraged...

This article discusses the creation of masks for highlighting lungs in computed tomography images using three methods – Otsu method, a simple convolutional neural network consisting 10 identical layers, and U-Net. We perform study comparison used automatically region interest (ROI) lungs, which were provided as courtesy from Clinics Samara State Medical University. The solution to this problem is relevant, because medical workers have manually select ROI first step automated processing lung...