Monochrome Images Research Articles

MONOCHROME AUGMENTED LOW-LIGHT IMAGE ENHANCEMENT

Low light short exposure photography is challenging, but an important factor in capturing images in temporarily dynamic scenes avoiding unwanted effects such as ghosting, motion blur, camera shakes, image artifacts, etc. Monochrome augmented low-light image enhancement aims to get improved low-light short-exposure images by using an additional monochrome sensor and its data. Monochrome images typically possess a higher SNR (Signal-to-Noise Ratio) and better luma information, since it avoids the attenuation by the Bayer Filter. The objective here is to develop a deep learning based approach to enhance low-light short exposure images from the main sensor by using an additional low resolution monochrome sensor.

Controlled Porosity of Selective Laser Melting-Produced Thermal Pipes: Experimental Analysis and Machine Learning Approach for Pore Recognition on Pipes Surfaces.

This study investigates the methods for controlling porosity in thermal pipes manufactured using selective laser melting (SLM) technology. Experiments conducted include water permeability tests and surface roughness measurements, which are complemented by SEM image ML-based analysis for pore recognition. The results elucidate the impact of SLM printing parameters on water permeability. Specifically, an increase in hatch and point distances leads to a linear rise in permeability, while higher laser power diminishes permeability. Using machine learning (ML) techniques, precise pore identification on SEM images depicting surface microstructures of the samples is achieved. The average percentage of the surface area containing detected pores for microstructure samples printed with laser parameters (laser power (W) _ hatch distance (µm) _ point distance (µm)) 175_ 80_80 was found to be 5.2%, while for 225_120_120, it was 4.2%, and for 275_160_160, it was 3.8%. Pore recognition was conducted using the Haar feature-based method, and the optimal patch size was determined to be 36 pixels on monochrome images of microstructures with a magnification of 33×, which were acquired using a Leica S9 D microscope.

Study on the quality of reproduction of graphical linear micro-images using electrophotographic printing

Today’s market of visual graphic communication includes many various advertising, publishing and packaging products printed on different materials. Many products are printed using conventional printing technologies such as offset printing or flexography, in addition to digital ones like Electrophotography and Ink Jet. The graphic images on prints often contain extremely fine graphic details, such as micro-lines or micro-text elements, which are part of various identification, security markings, codes, etc. It is a very important task to optimally model these micro-images on printed graphic communication products at the design stage, in order to ensure their brightness, legibility and geometric dimensions meet the requirements, not only immediately after printing, but also, for example, after a certain period during the life cycle of the printed product after exposure to the sunlight spectrum, or to other mechanical or environmental effects. This paper presents a methodology and results for the assessment of the quality of the reproduction of fine image details, i.e. the micro-lines of different widths arranged on the print in different directions. A specially developed original digital press quality control wedge with monochromatic micro-lines, microtext and screen dots bars of CMYK colours was applied to assess the accuracy of micro-image reproduction. In this part of the study, the wedge fields with positive micro-lines arranged in different directions individually and in groups were analysed. The width of the micro-lines was measured under the microscope to assess the accuracy of the dimensional reproduction and the deviations from the nominal value. To assess the influence of different parameters of the system “printing press-paper-inks” on the reproduction of image microelements, the prints for this part of the study were printed using electrophotographic printing techniques on different dry-toner printing presses on different types of paper. The width and deviations of the micro-lines on the print were measured along the print direction, perpendicular and at 45˚ to the print direction. Measurements of the reproduction accuracy of monochrome micro-line images, printed on dry-toner electrophotographic presses, showed that the accuracy of micro-line reproduction depends not only on the printing system, its resolution and the characteristics of the paper, but also on the direction of the positioning of the micro-lines on the printed sheet. The obtained results allow us to compare the capability of digital printing systems to reproduce linear micro-images on visual graphic communication products of any size and geometric orientation, to select optimal systems for printing specific products, and to model the layout of micro-images on products at the design stage, by assessing the orientation of the micro-images on the printed sheet.

Black and White Image Colorization with Deep Learning

Abstract: The project "Black and White Image Colorization with Deep Learning" develops a system to automatically infuse colors into grayscale images through cutting-edge deep learning approaches. Utilizing convolutional neural networks (CNNs) and generative adversarial networks (GANs), the system aims to deliver both realistic and visually pleasing colorization outcomes. This process includes training the deep learning models on diverse image datasets, optimizing the model parameters, and using advanced loss functions to ensure precise color representation.This project addresses the complex challenges associated with adding colors to monochrome images and also seeks to improve user experience by introducing customizable options such as style modification and color tweaks. By merging deep learning with computer vision techniques, this initiative is poised to contribute significantly to various image processing fields, especially in the restoration of historical images and the enhancement of visual content production. With thorough experimentation and ongoing refinements, the project aspires to establish a reliable and easy-to-use system that advances the state of image colorization technology in the modern digital age..

Quantitative skin surface hydration measurement by visible optical image processing: A pilot study.

Skin barrier function is significantly impacted by skin moisture. Most non-invasive evaluation techniques to measure skin surface hydration relying on its electrical properties, which are limited in scope and have unstable operations. Applying image processing for skin hydration assessment is uncommon, with an emphasis on skin-capacitive pictures and near-infrared images in general, which demand a certain spectrum. As a result, there is an increasing need for wide-area skin hydration evaluation and mapping. Thestudy aimsto propose a quantitative evaluation algorithmforskin surface hydrationfrom visible-light images. Threedeviceswere applied tomeasureskin hydration:skin image capturedevice and tworecognizedcommercialskindevices.A digital image processing system creates a new index, called GVR, to symbolize skin surface moisture. TheCLAHE algorithmwas appliedto enhance the contrast of skin image, andafter calculating it with the monochrome image,the skin reflectance imagewas segmented.The GVR was estimated using the values of the individual sites and the entire skin. The correlation coefficient between the three methods was examined using statistical analysis to assess the performance of GVR. Skin hydrationestimated from visible-light imagesis influenced by the entire facial structure in addition to specific areas.The electrical and visible image evaluations showed a strong association with a significant difference. It was discovered that reflecting measures from visible images provide a quick and efficient way to quantify the moisture of the skin's surface.

Dynamic range issues and image information content when visualizing seismic data

The purpose of this review is to discuss the problem of seismic data visualization in relation to human vision capabilities based on the seismic signal dynamic range assessment. The ways of increasing the information content and efficiency of visualization methods in terms of increasing their resolution are considered. The conducted analysis has shown that the dynamic range of seismic recordings exceeds the capabilities of the human eye by 2–3 orders of magnitude and the capabilities of technical visualization means by 4–5 orders of magnitude. Color vision models are considered on example of the most versatile hardware additive RGB color model. The model was chosen based on the idea of the human vision system as a set of photosensitive cones that register red, green and blue colors and the human brain, which processes and forms a color image in one’s consciousness. The relationships between monochrome and color images perceived by the human eye based on several types of sensitive receptors have been established. It is concluded that the use of a color image significantly expands information capabilities. Thus, 8 grey steps are enough for the monochrome method of information output, whereas the use of color devices allows to increase the number of steps up to 14. The methodology for solving this problem consists in the application of new technologies based on the use of large-scale and three-dimensional images providing a significant increase in the values of luminous flux, resolution and contrast based on the expansion of the dynamic range taking into consideration the limits of human vision.

Effect of Colorful Food Photography and Consumer Intentions to Purchase Food and Beverages

This study intends to look into how consumers are affected differently by food advertising utilizing monochrome and colorful images in terms of pleasure, arousal, and buy intention. A design study was carried out randomly with 101 participants from various residences in Jabodetabek Area (Jakarta, Bogor, Tangerang, Bekasi), Indonesia. Participants voluntarily exposed themselves to food advertisements. According to this study, employing either monochrome or colourful food photography can positively influence customer’s intentions to purchase foods and beverages. The results demonstrate that colourful food photography increases customer buy intentions by making food products more enticing and beautiful. One of the few research projects comparing the effect of colorful food photography as an advertising and consumer intentions to purchase food and beverages. The previous understanding of the significance of visuals in social media advertising is also expanded by this research.

Characteristics of the Retention of Color and Monochrome Images in Working Memory in the Monkey Macaca Mulatta

Characteristics of the Retention of Color and Monochrome Images in Working Memory in the Monkey Macaca Mulatta

Noise immunity algorithm for embedding a digital watermark in medical images

When storing and transferring printed medical materials, such as tomograms or radiographs, there is a need to protect additional information from unauthorized access. This information includes personal data of the patient and a summary of the medical history, and it can be added to the medical image (container) in the form of a watermark. Existing algorithms for embedding digital watermarks (DWM) in graphic objects distort the initial characteristics of the container image, which in the case of medical images can lead to a misdiagnosis. This study aimed to develop a distortion- and noise-resistant algorithm for embedding the DWM in the spatial domain of a medical image intended for storage on paper (as a printout). The article initially considered the possibilities of using the method of modifying the least significant bits of image pixel brightness or the LSB algorithm. Mathematical modeling in Matlab showed that the maximum brightness of the DWM guaranteeing its invisibility cannot exceed 2% of the maximum brightness of the container image for monochrome images, and 6% for color images. The maximum value of the white noise dispersion, at which it is possible to single out a DWM with a correlation coefficient of at least 0.9 was 0.0001. We prorose a new noise immunе algorithm (NIA) for embedding the DWM in the subpixels of the main image, which, after extracting the DWM, are not used to build the container image. In the absence of noise, there are no distortions of the original medical image whatsoever. The essence of the NIA is as follows. The size of the original image in the form of a two-dimensional array is quadrupled by adding a subpixel in each row and column with a brightness equal to the average arithmetic brightness of neighboring pixels. A DWM with the same size as the original image is added to the resulting subpixels. Matlab modeling showed that the DWM would remain invisible at a relative brightness of approximately 5% for monochrome container images and 15% for color images. For the NIA algorithm, the maximum value of the white noise dispersion to obtain a correlation coefficient of 0.9 is 0.005, which means that the noise immunity of the proposed method is significantly higher than that of the LSB-based algorithm.

Design and Development of Large-Band Dual-MSFA Sensor Camera for Precision Agriculture.

The optimal design and construction of multispectral cameras can remarkably reduce the costs of spectral imaging systems and efficiently decrease the amount of image processing and analysis required. Also, multispectral imaging provides effective imaging information through higher-resolution images. This study aimed to develop novel, multispectral cameras based on Fabry-Pérot technology for agricultural applications such as plant/weed separation, ripeness estimation, and disease detection. Two multispectral cameras were developed, covering visible and near-infrared ranges from 380 nm to 950 nm. A monochrome image sensor with a resolution of 1600 × 1200 pixels was used, and two multispectral filter arrays were developed and mounted on the sensors. The filter pitch was 4.5 μm, and each multispectral filter array consisted of eight bands. Band selection was performed using a genetic algorithm. For VIS and NIR filters, maximum RMS values of 0.0740 and 0.0986 were obtained, respectively. The spectral response of the filters in VIS was significant; however, in NIR, the spectral response of the filters after 830 nm decreased by half. In total, these cameras provided 16 spectral images in high resolution for agricultural purposes.

Relationships between colorization and pseudo-colorization of monochrome images

This paper investigates the relationship between colorization and pseudo-colorization techniques for converting grayscale images to color. Colorization strives to create visually believable color versions of monochrome images, either replicating the original colors or generating realistic, alternative color schemes. In contrast, pseudo-colorization maps grayscale intensities to pre-defined color palettes to improve visual appeal, enhance content understanding, or aid visual analysis. While colorization is an ill-posed problem with infinitely many RGB solutions, pseudo-colorization relies on mapping functions to deterministically assign colors. This work bridges these techniques by exploring the two following operations: first - deriving pseudo-color from colorized images - this allows for creating stylized or abstract representations from existing colorizations, and second - enriching color diversity in pseudo-colored images - this enhances visual appeal and attractiveness of pseudo-colored images. The paper emphasizes the centrality of decolorization (rgb-to-gray) models in both processes. It focuses on the theoretical underpinnings of these problems but complements them with illustrative examples for clarity.

Thermal-visible stereo matching at night based on Multi-Modal Autoencoder

Compared to RGB cameras, thermal sensors are more reliable at night. However, in the stereo matching task, it is difficult to estimate reliable depth simply depending on the binocular thermal cameras, due to their monochrome and low-contrast images. This paper proposed a thermal-visible stereo matching method based on multi-modal autoencoder (MANet), the key to which is to extract the modal-invariant features between thermal and visible images, and meanwhile, retain the characteristic features in terms of the optical properties. Besides, we designed a cross-reconstruction constraint to improve the scene feature extraction capability of MANet. In the stereo matching process, we analyze the thermal and visible properties and adopt optical attention mechanism to improve night-scene stereo matching. We established a stereo matching system with thermal-visible camera based on LiDAR calibration, and provided the dataset for evaluation. The experiment demonstrates that the proposed MANet can effectively extract the modal-invariant features and can be applied on thermal-visible stereo matching.

Exploring the Associations between Diabetes Mellitus and Diabetic Retinopathy: Prevention and Management by focus on Machine Learning Technique

Introduction: diabetes Mellitus, a disorder impacting insulin production and utilization, led to elevated blood sugar levels. Immune system assaults on insulin-producing pancreas cells caused Type 1 Diabetes Mellitus, while Type 2 Diabetes Mellitus affected glucose processing, predominantly in adults but also observed in children. Unmanaged diabetes resulted in varied health issues including heart disease, kidney damage, nerve impairment, and diabetic retinopathy, a major cause of adult blindness.Objective: to prevent diabetic retinopathy through glycemic control, achieved via management, lifestyle choices, screenings, treatments, education, and awareness. Machine learning techniques like transfer learning, ensemble learning, CNN-MNIST, and multiscale approaches showed promise in detection and diagnosis. Monitoring blood sugar and eye exams were vital for early retinopathy treatment.Result: DR risk is elevated in those with positive complications like nephropathy, heart disease, cerebrovascular disease, foot ulcers and HbA1C levels ≥6,8 %. Retinal imaging aids diagnosis and monitoring of ocular diseases like DR, utilizing processed monochrome images for structural analysis.Method: involved observing NPDR, MPDR via eye exams, measuring glucose, visual acuity, and retinal thickness. Retinal imaging aided ocular disease diagnosis, utilizing processed images for analysis.Conclusion: diabetes prevalence rose globally, projected to affect 800 million adults by 2050. High India rates emphasized healthcare need, especially in remote areas, addressing diabetic retinopathy and early symptom awareness

Raster image processing using 2D Padé-type approximations

We have developed a method called the two-dimensional Padé-type approximants method, which can be used to reduce the Gibbs phenomenon in the harmonic two-dimensional Fourier series. This method can be applied to both monochrome and color raster images. To do this, we implement the generalized two-dimensional Padé approximation proposed by Chisholm. In this approach, we select the range of frequency values on the integer grid according to the Vavilov method. We propose a definition of a Padé-type functional and provide examples of its application to simple discontinuous templates represented as raster images. Through this study, we are able to draw conclusions about the practical usage and advantages of the Padé-type approximation. We demonstrate that the Padé-type approximant effectively eliminates distortions associated with the Gibbs phenomenon, and it is visually more appropriate than the Fourier approximant. Additionally, the application of the Padé-type approximation reduces the number of parameters without sacrificing precision.

ML-aVAT: A Novel 2-Stage Machine-Learning Approach for Automatic Clustering Tendency Assessment

Clustering tendency assessment, which aims to deduce if a dataset contains any cluster structure, and, if it does, how many clusters it has, is a critical problem in exploratory data analysis. The VAT family of algorithms provides a “visual” means to assess the clustering tendency for various datasets. The VAT algorithm operates by reordering the pairwise distance matrix of the input data. When viewed as a monochrome image, this reordered dissimilarity matrix is called a reordered dissimilarity image (RDI), showing possible data clusters by dark blocks along the diagonal. This process, however, requires human intervention to interpret an RDI. Moreover, for datasets having complex cluster structure or noise, dark blocks along the diagonal of the RDI are not easily distinguishable, making it difficult to count them accurately, and different individuals can report different numbers of dark blocks. Only a handful of approaches have been proposed in the literature to automatically (algorithmically) infer the cluster structure from a VAT-type RDI without requiring human input. However, these approaches do not perform well for several data types and have impractically high run-time. This paper proposes and develops ML-aVAT: a novel two-stage machine-learning-based approach for automatic clustering tendency assessment from VAT-type RDI. Besides estimating the number of clusters, ML-aVAT can also infer the clustering hierarchy, i.e., sub-clusters within each group, something none of the previously proposed algorithms could do. Numerical experiments performed on various synthetic and real-life labeled and unlabeled datasets prove the effectiveness of ML-aVAT in estimating clustering tendency and cluster hierarchy.

Seasonal low albedo streaks at the edge of the Martian south polar layered deposits

Low Albedo Streaks (LAS) are narrow dark streaks observed at the edges of the South Polar Layered Deposits (SPLD) on Mars. We investigate the seasonal morphology, surface topography of LAS-rich areas using monochrome and multiband images from both the High Resolution Imaging Science Experiment (HiRISE) and the Color and Stereo Surface Imaging System (CaSSIS). We complement the image analysis with spectral data from the Compact Reconnassiance Imaging Spectrometer (CRISM). In addition, we analyze the thermal environment of LAS-rich areas with data from a Martian global climatological model. Our analysis suggests that streaks (i) originate from similar scarps and outcrops on the edge of layered structures at the end of local winter during different Martian years, (ii) lengthening in spring forming sinuous branches on terrain with rippled ridges, and (iii) eventually fade leaving distinguishable bright deposits within the streaked area. The surface topography analysis indicates that streaks advance on slopes between 4° and 34°, while the dust flows emanated from mass-wasting scars propagate on the slope steeper than 20°. Our spectral data analysis suggests that pyroxene- or basaltic-bearing materials are deposited on the surface mainly composed of dust by LAS. Thermal analysis indicates that LAS are active when the surface temperature rises from 145.5 K to 162 K, fading when the temperature increases suddenly and significantly. Finally, we evaluate several possible mechanisms to explain the formation of LAS. LAS might be formed by liquids, and such liquids might be brines. However, the presence of salts with lower eutectic temperatures seems to be necessary for the existence of brine on the Martian surface, and should be investigated at the target area in the future.

WORKING MEMORY OF <i>MACACA MULATTA</i> MONKEYS FOR COLOR AND MONOCHROME IMAGES

In behavioral experiments, we studied the ability of rhesus monkeys (Macaca mulatta) to keep in working memory visual objects that differ either in shape, or in color, or in a combination of these features. Six male rhesus monkeys performed a delayed matching-to-sample task, with three geometric shapes from a set of stimuli as samples. In the first series of experiments, these were colored figures of various shapes, in the second – circles of different colors, in the third – monochrome images of various figures from the set of stimuli. When using both features to memorize objects, the monkeys showed the maximum result, and the task of matching by color performed better than the task of matching by shape. The latter result disagrees with the data (Fehring et al., 2022), where in similar experiments, though with one sample, the opposite bias was observed. The reason for this may be the shift from local features (contours of shapes) to global ones (color) when recognizing and memorizing visual objects under conditions of a greater memory load in our study.

On a generalization of Tupper's formula for m colors and n dimensions

Tupper's formula 12<⌊mod(⌊y17⌋2−17⌊x⌋−mod(⌊y⌋,17),2)⌋ has an interesting property: that for any monochrome image that can be represented using a 106×17 array of two-dimensional pixels, there exists a natural number k such that the graph of the inequality over the range 0≤x<106 and k≤y<k+17 is that image. In this paper, we give a generalization for m colors and n dimensions. In this paper, by “natural numbers” we mean non-negative integers. We give m formulae employing n free variables with the property that, for any n-dimensional object of m colors C1,…,Cm that can be represented by hypervoxels (hypervoxels being the multidimensional analogue of pixels) in an n-dimensional array of dimensions A1×⋯×An, there exists a natural number k such that, when the first formula is graphed using color C1, the second formula is graphed using color C2, …, and the mth formula is graphed using color Cm over 0≤x1<A1, 0≤x2<A2,…,0≤xn−1<An−1, and k≤xn<k+An, the union of all colored graphs is that n-dimensional object.

Robust and Accurate Iris Segmentation Algorithm for Color and Noisy Eye Images

Efficient and robust segmentation of iris images captured in the uncontrolled environments is one of the challenges of non-cooperative iris recognition systems. We address this problem by proposing a novel iris segmentation algorithm, which is suitable both for monochrome and color eye images. The method presented use modified Hough transform to roughly localize the possible iris and pupil boundaries, approximating them by circles. A voting mechanisms is applied to select a candidate iris regions. The detailed iris boundary is approximated by the spline curve. Its shape is determined by minimizing introduced boundary energy function. The described algorithm was submitted to the NICE.I iris image segmentation contest, when it was ranked 11th and 10th out of total 97.

Backscattering Echo Intensity Characteristics of Laser in Soil Explosion Dust

Soil dust generated by explosions can lead to the absorption and scattering of lasers, resulting in low detection and recognition accuracy for laser-based devices. Field tests to assess laser transmission characteristics in soil explosion dust are dangerous and involve uncontrollable environmental conditions. Instead, we propose using high-speed cameras and an indoor explosion chamber to assess the backscattering echo intensity characteristics of lasers in dust generated by small-scale explosive blasts in soil. We analyzed the influence of the mass of the explosive, depth of burial, and soil moisture content on crater features and temporal and spatial distributions of soil explosion dust. We also measured the backscattering echo intensity of a 905 nm laser at different heights. The results showed that the concentration of soil explosion dust was highest in the first 500 ms. The minimum normalized peak echo voltage ranged from 0.318 to 0.658. The backscattering echo intensity of the laser was found to be strongly correlated with the mean gray value of the monochrome image of soil explosion dust. This study provides experimental data and a theoretical basis for the accurate detection and recognition of lasers in soil explosion dust environments.