Extraction of Hidden Science from Nanoscale Images DOI
Kristopher K. Barr, Naihao Chiang, Andrea L. Bertozzi

и другие.

The Journal of Physical Chemistry C, Год журнала: 2021, Номер 126(1), С. 3 - 13

Опубликована: Дек. 23, 2021

Scanning probe microscopies and spectroscopies enable investigation of surfaces even buried interfaces down to the scale chemical-bonding interactions, this capability has been enhanced with support computational algorithms for data acquisition image processing explore physical, chemical, biological phenomena. Here, we describe how scanning techniques have by some these recent algorithmic improvements. One improvement algorithm is advance beyond a simple rastering framework using spirals at constant angular velocity then switching linear velocity, which limits piezo creep hysteresis issues seen in traditional methods. can also use image-processing model distortions that appear from tip motion effects make corrections images. Another discuss enables researchers segment images domains subdomains, thereby highlighting reactive interesting disordered sites domain boundaries. Lastly, used examine dipole direction individual molecules surface domains, hydrogen bonding molecular tilt. The are still improving rapidly incorporating machine learning next level iteration. That said, not yet able perform live adjustments during recording could enhance microscopy spectroscopic imaging methods significantly.

Язык: Английский

A multi-modal image fusion framework based on guided filter and sparse representation DOI
Shuai Zhang, Fuyu Huang, Bingqi Liu

и другие.

Optics and Lasers in Engineering, Год журнала: 2020, Номер 137, С. 106354 - 106354

Опубликована: Авг. 26, 2020

Язык: Английский

Процитировано

30
Empirical curvelet transform based deep DenseNet model to predict NDVI using RGB drone imagery data DOI Creative Commons
Mohammed Diykh, Mumtaz Ali, Mehdi Jamei

и другие.

Computers and Electronics in Agriculture, Год журнала: 2024, Номер 221, С. 108964 - 108964

Опубликована: Апрель 23, 2024

Predicting accurately the Normalized Difference Vegetation Index (NDVI) trends from RGB images are essential to monitor crops and identify issues related plant diseases, water shortages. The current NDVI prediction models primarily based on traditional machine learning which lack reliability due problem atmospheric conditions. To predict in Prince Edward Island using drone imagery data, this paper proposed a novel framework integrating empirical curvelet transform DenseNet models. Each channel of data was passed through method where coefficients were analysed result creating new formula design NDVI. output sent deep final model evaluated quantitative metrics including, Q-Q plot, regression, correlation coefficients, structural similarity (SSIM), peak signal noise ratio (PSNR) mean square error (MSE) as well accuracy (ACC), sensitivity (SEN), f1-score, specificity. obtained results showed that outperformed previous by scoring highest values SSIM = 0.98, lowest MSE 120. It is believed helpful support farmers monitoring growth health problems.

Язык: Английский

Процитировано

4
A view of computational models for image segmentation DOI Creative Commons
Laura Antonelli, Valentina De Simone, Daniela di Serafino

и другие.

ANNALI DELL UNIVERSITA DI FERRARA, Год журнала: 2022, Номер 68(2), С. 277 - 294

Опубликована: Сен. 20, 2022

Abstract Image segmentation is a central topic in image processing and computer vision key issue many applications, e.g., medical imaging, microscopy, document analysis remote sensing. According to the human perception, process of dividing an into non-overlapping regions. These regions, which may correspond, different objects, are fundamental for correct interpretation classification scene represented by image. The division regions not unique, but it depends on application, i.e., must be driven final goal hence most significant features with respect that goal. Thus, can regarded as strongly ill-posed problem. A classical approach deal ill posedness consists incorporating model a-priori information about solution, form penalty terms. In this work we provide brief overview basic computational models segmentation, focusing edge-based region-based variational models, well statistical machine-learning approaches. We also sketch numerical methods applied computing solutions these models. our opinion, view help readers identify suitable classes solving their specific problems.

Язык: Английский

Процитировано

9
Nano-scale wear: A critical review on its measuring methods and parameters affecting nano-tribology DOI
Behzad Sadeghi, Pasquale Cavaliere, Ali Shabani

и другие.

Proceedings of the Institution of Mechanical Engineers Part J Journal of Engineering Tribology, Год журнала: 2023, Номер 238(2), С. 125 - 155

Опубликована: Ноя. 7, 2023

This paper presents a critical review on the measuring methods and parameters affecting nano-tribology in context of nano-scale wear. Nano-scale wear phenomena play crucial role various industries, including micro/nano-systems, electronics, biotechnology. The begins by discussing significance its impact device performance, lifespan, durability, energy efficiency, cost savings, environmental sustainability. It then delves into employed to assess wear, scanning probe microscopy (SPM) techniques such as atomic force (AFM) friction (FFM). capabilities AFM FFM studying roughness surface, adhesion, friction, scratch, abrasion, material transfer are highlighted. Additionally, explores nano-wear, lubrication strategies, stress levels, sliding velocity, atomic-scale reactions. article concludes emphasizing importance advanced understanding tribological mechanisms at different scales, bridging gap between macro studies.

Язык: Английский

Процитировано

5
Demon Registration for 2D Empirical Wavelet Transforms DOI Creative Commons
Charles-Gérard Lucas, Jérôme Gilles

Foundations, Год журнала: 2024, Номер 4(4), С. 690 - 703

Опубликована: Дек. 3, 2024

The empirical wavelet transform is a fully adaptive time-scale representation that has been widely used in the last decade. Inspired by mode decomposition, it consists of filter banks based on harmonic supports. Recently, generalized to build from any generating function using mappings. In practice, supports can have low-constrained shape 2D, leading numerical difficulties estimate mappings adapted construction filters. This work aims propose an efficient scheme compute coefficients demons registration algorithm. Results show proposed approach robust, accurate, and continuous filters permitting reconstruction with low signal-to-noise ratio. An application for texture segmentation scanning tunneling microscope images also presented.

Язык: Английский

Процитировано

1
Difference of anisotropic and isotropic TV for segmentation under blur and Poisson noise DOI Creative Commons
Kevin Bui, Yifei Lou,

Fredrick Park

и другие.

Frontiers in Computer Science, Год журнала: 2023, Номер 5

Опубликована: Июнь 28, 2023

In this paper, we aim to segment an image degraded by blur and Poisson noise. We adopt a smoothing-and-thresholding (SaT) segmentation framework that finds piecewise-smooth solution, followed k -means clustering the image. Specifically for smoothing step, replace least-squares fidelity Gaussian noise in Mumford-Shah model with maximum posterior (MAP) term deal incorporate weighted difference of anisotropic isotropic total variation (AITV) as regularization promote sparsity gradients. For such nonconvex model, develop specific splitting scheme utilize proximal operator apply alternating direction method multipliers (ADMM). Convergence analysis is provided validate efficacy ADMM scheme. Numerical experiments on various scenarios (grayscale/color multiphase) showcase our proposed outperforms number methods, including original SaT.

Язык: Английский

Процитировано

2
Particle and Particle Agglomerate Size Monitoring by Scanning Probe Microscope DOI Creative Commons
P. V. Gulyaev, Tibor Krenický,

E. Yu. Shelkovnikov

и другие.

Applied Sciences, Год журнала: 2022, Номер 12(4), С. 2183 - 2183

Опубликована: Фев. 19, 2022

In the present study, use of a scanning probe microscope is described for monitoring sizes nanoparticles. Monitoring process acquiring and analysing set overlapping images. The main analysis steps are image segmentation, determination nanoparticles allocation their sizes, overlap images with one another, exclusion repeating measurements formation correct particle-size sampling. thorough examination commercial microscopes, software, processing libraries showed that capabilities limited segmentation in complex structured A method based on surface curvature computation proposed (allocation particles) particle sizes. estimated using area approximation respect to circumference. It sample displacement sensors as an aid stitching.

Язык: Английский

Процитировано

3
Application of neural network of U-Net architecture for segmentation of nanoparticles on CTM-probes DOI Open Access

E Shelkovnikov,

K. A. Shlyakhtin,

Tatyana Shelkovnikova

и другие.

Himičeskaâ fizika i mezoskopiâ, Год журнала: 2019, Номер 21(2)

Опубликована: Июнь 15, 2019

Язык: Английский

Процитировано

3
Application of the Hough Transform to Dispersion Control of Overlapping Particles and Their Agglomerates DOI Creative Commons
P. V. Gulyaev

Devices and Methods of Measurements, Год журнала: 2023, Номер 14(3), С. 199 - 206

Опубликована: Окт. 6, 2023

The dispersion control of micro- and nanoparticles by their images is great importance for ensuring the specified properties particles themselves materials based on them. aim this article was to consider possibilities using Hough transform overlapping agglomerates. Analysis application agglomerates showed following. particularities conventional implementation lead preferred registration large particles, shift centers distortion size values. To use correctly, fine-tuning all its parameters required. automate process, dependences number recorded in image investigated. studies were carried out test with a known particles. results that when threshold change, detected stabilizes near optimal When range changes, histogram particle distribution changes. In case, width determined most stable extremes histogram. maximum center-to-center distance set at least half range. configuration algorithm described implemented. It implies repeatedly running different combinations parameters. includes stages coarse fine-tuning, which allows getting closer efficiency has been confirmed real images. Tests have shown errors determining multi-pass are same level or exceed these indicators analog methods.

Язык: Английский

Процитировано

1
Empirical Voronoi wavelets DOI Creative Commons
Jérôme Gilles

Constructive Mathematical Analysis, Год журнала: 2022, Номер 5(4), С. 183 - 189

Опубликована: Ноя. 1, 2022

Recently, the construction of 2D empirical wavelets based on partitioning Fourier domain with watershed transform has been proposed. If such approach can build partitions completely arbitrary shapes, for some applications, it is desirable to keep a certain level regularity in geometry obtained partitions. In this paper, we propose partition using Voronoi diagrams. This solution allows us high adaptability while guaranteeing minimum geometric detected partition.

Язык: Английский

Процитировано

1