Binary Input Image Research Articles

Text Localization and Enhancement of Mobile Camera based Complex Natural Bilingual Text Scene Images

In this paper, we proposed a text localization and enhancement of the Mobile camera based complex natural scene images using fixed coordinates for removal of unwanted non text regions and making it localized. The morphological erosion technique is applied for the smoothness of foreground text regions. Here, the experimental working nature of the proposed method is first localized color input image is converted into a gray level, then a median filter is applied for the removal of noises. After that, the input image is further converted into binary image. In addition to this, morphological dilation and erosion operations also applied for connecting the broken pixels. Further, the connected component segmentation technique is applied to extract the foreground text regions from the binary input image. The real time input samples are collected by using mobile camera from the various categories of natural scenes of monolingual/bilingual text such as Stone, Wall, Poster and Iron-based. A total of 140 real time samples of natural scene text images are considered for the experimental setup. The experimental results shown that 132 real time sample images are localized and then obtained the accuracy of 94.24%. Further, bilingual text regions are enhanced using hybrid approach and obtain the segmentation accuracy of 83.17%. The novelty of the paper is working on the real time data set and it contains a mixture of printed and handwritten natural scene text images.

Fast Grayscale Morphology for Circular Window

AbstractMorphological operations are among the most popular classic image filters. The filter assumes the maximum or minimum value within a window and is often used for light object thickening and thinning operations, which are important components of various workflows, such as object recognition and stylization. Circular windows are preferred over rectangular windows for obtaining isotropic filter results. However, the existing efficient algorithms focus on rectangular or binary input images. Efficient morphological operations with circular windows for grayscale images remain challenging. In this study, we present a fast grayscale morphology heuristic computation algorithm that decomposes circular windows using the convex hull of circles. We significantly accelerate traditional methods based on Minkowski addition by introducing new decomposition rules specialized for circular windows. As our morphological operation using a convex hull can be computed independently for each pixel, the algorithm is efficient for modern multithreaded hardware.

Deep learning particle identification in LHCb RICH

The use of Ring Imaging Cherenkov detectors (RICH) offers a powerful technique for identifying the particle species in particle physics. These detectors produce 2D images formed by rings of individual photons superimposed on a background of photon rings from other particles.The RICH particle identification (PID) is essential to the LHCb experiment at CERN. While the current PID algorithm has performed well during LHC data-taking periods between 2010 to 2018, its complexity poses a challenge for LHCb computing infrastructure upgrades towards multi-core architectures. The high particle multiplicity environment of future LHC runs strongly motivates shifting towards high-throughput computing for the online event reconstruction.In this contribution, we introduce a convolutional neural network (CNN) approach to particle identification in LHCb RICH. The CNN takes binary input images from the two RICH detectors to classify particle species. The input images are polar-transformed sub-sections of the RICH photon-detection planes. The model is hyperparameter-optimised and trained on classification accuracy with simulated collision data for the upcoming LHC operation starting in 2022. The PID performance of the CNN is comparable to the conventional algorithm, and its simplicity renders it suitable for fast online reconstruction through parallel processing.We show that under conditions of reduced combinatorial background, as expected from the introduction of timing resolution to the RICH detectors in future upgrades, the network achieves a particle identification performance close to 100 %, with simultaneous misclassification of the most prevalent particle species approaching 0 %.

A technical note on digitizing color mapped spectral power distribution images

AbstractData visualization is producing images that communicate data in the form of visual objects like lines, points, bars, or colored areas. Often it is necessary to extract numerical data from these images for further analysis. There are a wide variety of data digitization tools, however, only limited formats of plot images can be digitized using them. There exists many data visualization formats spread across different domains of science. Some of these formats need tailored solutions. One such format is color mapped spectral power distribution images of light sources captured by proprietary spectrometers. In this work, a methodology is discussed to digitize spectral plots of light sources. Unlike typical digitization tools which extract data based on contrast difference, color, or morphological similarities, color mapped spectral plots contain blending colors across the x‐axis, without containing any notable morphological features. In this work, image processing is applied to extract the contour representing a spectral plot. For color mapped spectral plots, the importance of appropriate binary input images for edge detection algorithms is highlighted. There are only a few edge‐detection methods give the desired results except a few. Further, the importance of repeatedly adjusting threshold values in extracting desired contours is also discussed. The estimated and measured spectral parameters are in mutual agreement with each other thereby validating the adopted approach.

Seeing through multimode fibers with real-valued intensity transmission matrices

Image transmission through multimode optical fibers has been an area of immense interests driven by the demand for miniature endoscopes in biomedicine and higher speed and capacity in telecommunications. Conventionally, a complex-valued transmission matrix is obtained experimentally to link the input and output light fields of a multimode fiber for image retrieval, which complicates the experimental setup and increases the computational complexity. Here, we report a simple and high-speed method for image retrieval based on our demonstration of a pseudo-linearity between the input and output light intensity distributions of multimode fibers. We studied the impact of several key parameters to image retrieval, including image pixel count, fiber core diameter and numerical aperture. We further demonstrated with experiments and numerical simulations that a wide variety of input binary and gray scale images could be faithfully retrieved from the corresponding output speckle patterns. Thus, it promises to be useful for highly miniaturized endoscopy in biomedicine and spatial-mode-division multiplexing in telecommunications.

A Parallel Connected Component Labeling Architecture for Heterogeneous Systems-on-Chip

Connected component labeling is one of the most important processes for image analysis, image understanding, pattern recognition, and computer vision. It performs inherently sequential operations to scan a binary input image and to assign a unique label to all pixels of each object. This paper presents a novel hardware-oriented labeling approach able to process input pixels in parallel, thus speeding up the labeling task with respect to state-of-the-art competitors. For purposes of comparison with existing designs, several hardware implementations are characterized for different image sizes and realization platforms. The obtained results demonstrate that frame rates and resource efficiency significantly higher than existing counterparts are achieved. The proposed hardware architecture is purposely designed to comply with the fourth generation of the advanced extensible interface (AXI4) protocol and to store intermediate and final outputs within an off-chip memory. Therefore, it can be directly integrated as a custom accelerator in virtually any modern heterogeneous embedded system-on-chip (SoC). As an example, when integrated within the Xilinx Zynq-7000 X C7Z020 SoC, the novel design processes more than 1.9 pixels per clock cycle, thus furnishing more than 30 2k × 2k labeled frames per second by using 3688 Look-Up Tables (LUTs), 1415 Flip Flops (FFs), and 10 kb of on-chip memory.

An Unsupervised Retinal Vessel Segmentation Using Hessian and Intensity Based Approach

The structure of blood vessels play a crucial role in diagnoses of the various vision threatening diseases including Glaucoma and Diabetic Retinopathy (DR). The correct segmentation of retinal blood vessels is a crucial step in the study of retinal fundus images. We proposed a simple unsupervised approach by using a combination of Hessian based approach and intensity transformation approach. We have applied CLAHE for enhancing the contrast of the retinal fundus images. An enhanced version of PSO algorithm is applied for contextual region tuning of CLAHE. Morphological filter and Wiener filter are used to de-noise the enhanced image. The eigenvalues are obtained from the Hessian matrix at two different scales to extract thick and thin vessel enhanced images separately. The intensity transformation approach is separately applied to the enhanced image to maximize the vessel details. Global Otsu thresholding is applied on intensity transformed image and thick vessel enhanced image whereas ISODATA local thresholding is applied on thin vessel enhanced image. Finally, a simple post-processing step based on the region parameters such as area, eccentricity, and solidity is used. The region parameters are obtained for each connected component in input binary images. The threshold values of region parameters are empirically investigated and applied to each of the three binary images to remove the non-vessel components. The thresholded images are combined by applying logical OR operator, which resulted in the final segmented binary image. We assessed our developed framework on the open-access CHASE_DB1 and DRIVE datasets, achieving a sensitivity of 0.7776 and 0.7851, and an accuracy of 0.9505 and 0.9559 respectively. These results outperform several state-of-the-art unsupervised methods. The reduced computational complexity and significantly improved evaluation metrics advocates for its use in the automated diagnostic systems for retinal image analysis.

Performance Analysis of ANN based Gait Recognition

It is well-known that biometrics are a powerful tool for reliable automated person identification. Automatic gait recognition is one of the newest of the emergent biometrics and has many advantages over other biometrics. The most notable advantage is that it does not require contact with the subjects nor does it require the subject to be near a camera. This work employs a gait recognition process with binary silhouette-based input images and Artificial Neural Network (ANN) based classification in MATLAB. The performance of the recognition method depends significantly on the quality of the extracted binary silhouettes. In this work, a computationally low-cost fuzzy correlogram based method is employed for background subtraction. Even highly robust background subtraction and shadow elimination algorithms produce erroneous outputs at times with missing body portions, which consequently affect the recognition performance. Frame Difference Energy Image (FDEI) reconstruction is performed to alleviate the detrimental effect of improperly extracted silhouettes and to make the recognition method robust to partial incompleteness. Subsequently, features are extracted via two methods and fed to the BPNN (Back Propagation Neural Network based classifier which uses feature vector (exemplars) to compute similarity scores and carry out identification using weight vectors i.e. Frame-to-Exemplar-Distance (FED) vector. The FED uses the distance measure between pre-determined feature vectors and the weight vectors of the current frame as an identification criterion. The ANN performance is evaluated for recognition and speed parameters at different training gait angles.

ВИБІР НАЛАШТУВАНЬ ПІД ЧАС ОБЧИСЛЕННЯ ПОЛЯ ФРАКТАЛЬНИХ РОЗМІРНОСТЕЙ ЗОБРАЖЕННЯ

Результати досліджень в біології, медицині, дистанційному зондуванні Землі, матеріалознавстві доволі часто представляють у вигляді цифрових зображень. Такі зображення, здебільшого, характеризуються не дуже високою візуальною якістю та складною структурою. Усі ці чинники істотно впливають на подальший аналіз візуальних даних. Незважаючи на інтелектуальні властивості зорового сприйняття, людина не може справитися з великим потоком інформації та забезпечити необхідну продуктивність її опрацювання, тому кількісний аналіз візуальних даних доцільно перекласти на обчислювальні системи. Це призвело до того, що останнім часом розвиток значного сегменту сучасних інформаційних технологій націлений на створення методів опрацювання та аналізу візуальної інформації. Результат процесу опрацювання та аналізу є надзвичайно важливим, оскільки він лежить в основі прийняття відповідальних рішень, для прикладу, в медицині чи на виробництві. Традиційні підходи, які базуються на евклідовій геометрії, не завжди є ефективними в разі опрацювання зображень зі складною структурою. Перспективним інструментарієм для аналізу таких зображень є фрактальна розмірність. Загалом вона дає змогу обчислювати тільки глобальні характеристики зображення. Для кількісного аналізу складноструктурованих зображень доцільно формувати поле фрактальних розмірностей. Встановлено, що для обчислення поля фрактальних розмірностей важливим є правильний вибір налаштувань. Серед них треба виділити – розмір локальної апертури, крок зміщення та метод бінаризації. Розмір локальної апертури істотно впливає на деталізацію під час аналізу зображення. Для найкращої локалізації розмір ковзного околу повинен бути співмірним з розміром об'єктів зацікавлення на зображенні. Другим важливим параметром є крок зміщення локальної апертури. Він пропорційний до часу опрацювання та обернено пропорційний до розміру зображення поля фрактальних розмірностей. Збільшення кроку зміщення доцільне під час опрацювання зображень великих розмірів, але це призводить до зменшення деталізації об'єктів. Оскільки більшість методів обчислення фрактальних розмірностей використовує за вхідні бінарні зображення, важливим є вибір методу порогової бінаризації, який має забезпечити найкращу деталізацію об'єктів інтересу.

Image quality enhancement and computation acceleration of 3D holographic display using a symmetrical 3D GS algorithm.

The 3D Gerchberg-Saxton (GS) algorithm can be used to compute a computer-generated hologram (CGH) to produce a 3D holographic display. But, using the 3D GS method, there exists a serious distortion in reconstructions of binary input images. We have eliminated the distortion and improved the image quality of the reconstructions by a maximum of 486%, using a symmetrical 3D GS algorithm that is developed based on a traditional 3D GS algorithm. In addition, the hologram computation speed has been accelerated by 9.28 times, which is significant for real-time holographic displays.

Reconfigurable Processor for Binary Image Processing

Binary image processing is a powerful tool in many image and video applications. A reconfigurable processor is presented for binary image processing in this paper. The processor's architecture is a combination of a reconfigurable binary processing module, input and output image control units, and peripheral circuits. The reconfigurable binary processing module, which consists of mixed-grained reconfigurable binary compute units and output control logic, performs binary image processing operations, especially mathematical morphology operations, and implements related algorithms more than 200 f/s for a 1024 × 1024 image. The periphery circuits control the whole image processing and dynamic reconfiguration process. The processor is implemented on an EP2S180 field-programmable gate array. Synthesis results show that the presented processor can deliver 60.72 GOPS and 23.72 GOPS/mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> at a 220-MHz system clock in the SMIC 0.18-μm CMOS process. The simulation and experimental results demonstrate that the processor is suitable for real-time binary image processing applications.

Application of direct tracking method for measuring electrospun nanofiber diameter

In this paper, direct tracking method as an image analysis based technique for measuring electrospun nanofiber diameter has been presented and compared with distance transform method. Samples with known characteristics generated using a simulation scheme known as µ-randomness were employed to evaluate the accuracy of the method. Electrospun webs of polyvinyl alcohol (PVA) were also used to verify the applicability of the method on real samples. Since direct tracking as well as distance transform require binary input images, micrographs of the electrospun webs obtained from Scanning Electron Microscopy (SEM) were first converted to black and white using local thresholding. Direct tracking resulted in more accurate estimations of fiber diameter for simulated images as well as electrospun webs suggesting the usefulness of the method for electrospun nanofiber diameter measurement.

Improvement in holographic storage capacity by use of double-random phase encryption

We show that a double-random encryption technique can improve the storage capacity of an angular-multiplexed holographic memory system. In the holographic memory system, input binary images are encrypted into white-noise-like images by use of two random phase masks located at the input and the Fourier planes. These encrypted images are stored as holograms in a photorefractive medium by use of angular multiplexing. All the images are encrypted by different sets of random phase masks. Even when the angle separation between adjacent images is small enough to cause cross talk between adjacent images, original binary data can be recovered with the correct phase mask; the other reconstructed images remain white-noise-like images because incorrect masks are used. Therefore the capacity of the proposed system can be larger than that of a conventional holographic memory system without the random phase encryption technique. Numerical evaluation and experimental results are presented to confirm that the capacity of the system with random phase masks is larger than that of the conventional memory system.

Photorefractive multichannel correlator based on stoichiometric LiTaO3

We demonstrate a Vander Lugt type correlator based on photorefractive stoichiometric LiTaO3 as the volume holographic material. Accurate correlator output was obtained for binary phase-modulated input images. The correlator is tested with binary amplitude- and phase-modulated input images. Phase-modulated images show a high-quality correlation, whereas amplitude-modulated images suffer in part from a photorefractive self-focusing effect. Using 100 previously stored holograms as filters, a rate of 10000 correlations/s are demonstrated. A paraxial theoretical analysis of the correlator, giving analytic expressions for the shift invariance, is also described in detail.

Comparative effects of optical-correlator signal-dependent and signal-independent noise on pattern-recognition performance with the phase-only filter

The comparative effects of optical-correlator signal-dependent and additive signal-independent noise on correlation-filter performance are analyzed by three different performance measures. For an identical value of the signal-to-noise ratio imposed on each type of noise in a binary input image, computer simulations performed with the phase-only filter show (i) that additive Gaussian signal-independent noise yields a much larger correlation-performance degradation than signal-dependent noise and (ii) that the different types of signal-dependent noise lead to similar correlation results because of similar effects on the input image that are inherent to the nature of the noise.

Binary image synthesis using mixed linear integer programming

Images that are to be transmitted through a distorting imaging system may be deliberately altered to compensate for that distortion. The authors consider an incoherent diffraction-limited imaging system followed by an ideal high-contrast detector that prints binary images, and seek a binary input image (a mask) that generates a desired prescribed binary output image. They have developed a method for determining the optimal binary mask by formulating the problem as a mixed linear integer program (MLIP) and using the branch and bound method to solve it.

Optical asymmetrical median filtering using gray-scale convolution kernels

Morphological transformations such as erosion and dilation can be performed optically by convolving the binary input image with a binary kernel and then passing the result through an ideal thresholder-hardlimiter. Generalizations result from using gray-scale convolution kernels. It is shown that gray-scale kernels can be used to perform transformations similar to openings and closing with just one convolution instead of the ususal cascade of erosion and dilation normally required for these operations. Optical experimental results are presented.

Threshold superposition in morphological image analysis systems

It is shown that four composite morphological systems, namely morphological edge detection, peak/valley extraction, skeletonization, and shape-size distributions obey a weak linear superposition, called threshold-linear superposition. The output image signal or measurement from each system is shown to be the sum of outputs due to input binary images that result from thresholding the input gray-level image at all levels. These results are generalized to a vector space formulation, e.g. to any finite linear combination of simple morphological systems. Thus many such systems processing gray-level images are reduced to corresponding binary image processing systems, which are easier to analyze and implement.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>