Volume Of Image Data Research Articles

Dynamic Ensemble Learning with Evolutionary Programming and Swarm Intelligence for Image Classification

Image classification is a fundamental and pervasive task in the field of computer vision, with profound implications across a wide array of applications. From the autonomous vehicles navigating complex urban environments to the critical diagnosis of diseases through medical image analysis, the accurate categorization of images plays a pivotal role in modern technology and society. As the world becomes increasingly digitized, the volume of image data generated daily continues to soar. This exponential growth necessitates the development of robust and adaptive image classification techniques capable of handling the complexity and diversity of this data.In response to these evolving demands, this work introduces a pioneering approach known as Dynamic Ensemble Learning with Evolutionary Programming and Swarm Intelligence (DEL-EPSI). DEL-EPSI represents a transformative paradigm shift in the realm of image classification, offering a multifaceted solution to address the multifarious challenges posed by this critical task. At its core, DEL-EPSI combines the strengths of dynamic ensemble learning, evolutionary programming, and swarm intelligence to provide a holistic, adaptive, and highly accurate image classification framework. In this study, the principles, capabilities, and implications of DEL-EPSI are studied and investigated.

DCE-MRI and DWI can differentiate benign from malignant prostate tumors when serum PSA is ≥10 ng/ml.

This study investigated the diagnostic utility of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and diffusion-weighted imaging (DWI) parameters for distinguishing between benign and malignant prostate tumors when serum prostate-specific antigen (PSA) level is ≥10 ng/ml. Patients with prostate cancer (PCa) and benign prostatic hyperplasia (BPH) with serum PSA ≥10 ng/ml before treatment were recruited. Transrectal ultrasound-guided biopsy or surgery was performed for tumor classification and patients were stratified accordingly into PCa and BPH groups. Patients underwent DCE-MRI and DWI scanning and the transfer constant (Ktrans), rate constant (Kep), fractional volume of the extravascular extracellular space, plasma volume (Vp), and Prostate Imaging Reporting and Data System Version 2 (PI-RADS v2) score were determined. The apparent diffusion coefficient (ADC) was calculated from DWI. The diagnostic performance of these parameters was assessed by receiver operating characteristic (ROC) curve analysis, and those showing a significant difference between the PCa and BPH groups were combined into a multivariate logistic regression model for PCa diagnosis. Spearman's correlation was used to analyze the relationship between Gleason score and imaging parameters. The study enrolled 65 patients including 32 with PCa and 33 with BPH. Ktrans (P=0.006), Kep (P=0.001), and Vp (P=0.009) from DCE-MRI and ADC (P<0.001) from DWI could distinguish between the 2 groups when PSA was ≥10 ng/ml. PI-RADS score (area under the ROC curve [AUC]=0.705), Ktrans (AUC=0.700), Kep (AUC=0.737), Vp (AUC=0.688), and ADC (AUC=0.999) showed high diagnostic performance for discriminating PCa from BPH. A combined model based on PI-RADS score, Ktrans, Kep, Vp, and ADC had a higher AUC (1.000), with a sensitivity of 0.998 and specificity of 0.999. Imaging markers showed no significant correlation with Gleason score in PCa. DCE-MRI and DWI parameters can distinguish between benign and malignant prostate tumors in patients with serum PSA ≥10 ng/ml.

An Efficient Architecture and High-Throughput Implementation of CCSDS-123.0-B-2 Hybrid Entropy Coder Targeting Space-Grade SRAM FPGA Technology

Nowadays, hyperspectral imaging is recognized as cornerstone remote sensing technology. The explosive growth in image data volume and instrument data rates, compete with limited on-board storage resources and downlink bandwidth, making hyperspectral image data compression a mission critical on-board processing task. The Consultative Committee for Space Data Systems (CCSDS) extended the previous issue of the CCSDS-123.0 Recommended Standard for multi- and hyperspectral image compression to provide with Near-Lossless compression functionality. A key feature of the CCSDS-123.0-B-2 is the improved Hybrid Entropy Coder, which at low bit rates, provides substantially better compression performance than the Issue 1 entropy coders. In this paper, we introduce a high-throughput hardware implementation of the CCSDS-123.0-B-2 Hybrid Entropy Coder. The introduced architecture exploits the systolic design pattern to provide modularity and latency insensitivity in a deep and elastic pipeline achieving a constant throughput of 1 sample/cycle with a small FPGA resource footprint. This architecture is described in portable VHDL RTL and is implemented, validated and demonstrated on a commercially available Xilinx KCU105 development board hosting a Xilinx Kintex Ultrascale XCKU040 SRAM FPGA, and thus, is directly transferable to Xilinx Radiation Tolerant Kintex UltraScale XQRKU060 space-grade devices for space deployments. Moreover, state-of-the-art SpaceFibre (ECSS-E-ST-50-11C) serial link interface and test equipment were used in the validation platform to emulate an on-board deployment. The introduced CCSDS-123.0-B-2 Hybrid Entropy Encoder achieves a constant throughput performance of 305 MSamples/s. To the best of our knowledge, this is the first published fully-compliant architecture and high-throughput implementation of the CCSDS-123.0-B-2 Hybrid Entropy Coder, targeting space-grade FPGA technology.

Reduction in required volume of imaging data and image reconstruction time for adaptive-illumination Fourier ptychographic microscopy.

Fourier ptychographic microscopy (FPM) is a promising super-resolution computational imaging technology. It stitches a series of low-resolution (LR) images in the Fourier domain by an iterative method. Thus, it obtains a large field of view and high-resolution quantitative phase images. Owing to its capability to perform high-spatial bandwidth product imaging, FPM is widely used in the reconstruction of conventional static samples. However, the influence of the FPM imaging mechanism limits its application in high-speed dynamic imaging. To solve this problem, an adaptive-illumination FPM scheme using regional energy estimation is proposed. Starting with several captured real LR images, the energy distribution of all LR images is estimated, and select the measurement images with large information to perform FPM reconstruction. Simulation and experimental results show that the method produces efficient imaging performance and reduces the required volume of data to more than 65% while ensuring the quality of FPM reconstruction.

Fully Convolutional Network for the Semantic Segmentation of Medical Images: A Survey.

There have been major developments in deep learning in computer vision since the 2010s. Deep learning has contributed to a wealth of data in medical image processing, and semantic segmentation is a salient technique in this field. This study retrospectively reviews recent studies on the application of deep learning for segmentation tasks in medical imaging and proposes potential directions for future development, including model development, data augmentation processing, and dataset creation. The strengths and deficiencies of studies on models and data augmentation, as well as their application to medical image segmentation, were analyzed. Fully convolutional network developments have led to the creation of the U-Net and its derivatives. Another noteworthy image segmentation model is DeepLab. Regarding data augmentation, due to the low data volume of medical images, most studies focus on means to increase the wealth of medical image data. Generative adversarial networks (GAN) increase data volume via deep learning. Despite the increasing types of medical image datasets, there is still a deficiency of datasets on specific problems, which should be improved moving forward. Considering the wealth of ongoing research on the application of deep learning processing to medical image segmentation, the data volume and practical clinical application problems must be addressed to ensure that the results are properly applied.

A novel method of data compression using ROI for biomedical 2D images

In the era of modern medical imaging communal, huge volume of medical image data are being acquired and need to be transmitted and archived which necessitate the development of efficient image compression techniques on both 2D and 3D images. The compression of medical images is an essential process to support remote healthcare services. The medical diagnostics through these services require more accurate information from an image. As the property of inverse proportionality between the compression rate and quality of the image takes place in any kind of compression method, there is a need to sacrifice any one of those credentials (Quality or Compression Rate). With this context, Region of Interest (ROI) codecs are emerging and reduces this proportionality that yields more compression rate without compromising the quality. In this paper, presents an ROI based near lossless image compression method that incorporates the Set Partitioning in Hierarchical Trees (SPIHT) and Vector Quantization coding for medical images.

3D visualization of bioerosion in archaeological bone

Palaeoradiology is increasingly being used in archaeological and forensic sciences as a minimally invasive alternative to traditional histological methods for investigating bone microanatomy and its destruction by diagenetic processes. To better understand ancient mortuary practices, taphonomic studies using microCT scanning methods are gaining an ever more important role. Recently it was demonstrated that 2D virtual sections obtained by microCT scanning of intact samples are comparable to physical sections for the rating and diagnosis of bioerosion in archaeological bone. Importantly, volume image data obtained from tomographic methods also allow the rendering and analysis of 3D models. Building on these methods we provide (1) detailed descriptions of bioerosion in 3D volume renderings, virtual sections, and traditional micrographs, and (2) accessible techniques for the visualization of bioerosion in skeletal samples. The dataset is based on twenty-eight cortical bone samples, including twenty femora (of which five are cremated), two ribs, two parietals, one mandibular ramus, one humerus, and two faunal long bones from five archaeological sites in Lower Austria dating from the Early Neolithic to the Late Iron Age. Notably, we reduce the need for time-consuming image segmentation by sequentially applying two noise-reducing, edge-preserving filters, and using an image-display transfer function that visualizes bioerosion, as well as Haversian and Volkmann canal structure and density in 3D. In doing so we are also able to visualize in 3D the invasion of canals by microbiota, which has previously only been reported in 2D sections. Unlike conventional thin sections, the 3D volume images shown here are easy to create and interpret, even for archaeologists inexperienced in histology, and readily facilitate the illustration and communication of microtaphonomic effects.

An Efficient Compression of Gray Scale Images Using Wavelet Transform

In the field of digital image distribution and reproduction, the standardization of digital photography and the recent advancement in technology have led to explosive growth. This new impetus has increased the need for introducing scientifically, well designed image compression methods that would help to handle, store and transmit a huge volume of image data. Today, digital image compression is an important role in the fields of multimedia. DWT in corporation with threshold and quadtree decomposition of absorb new designed algorithm is the objective of the current study. In respect of quality execution of this design, it is observed in parity of the EZW image compression technique at the similar bit rating, however, any other habitual standard image does not encourage compression method.

Three-dimensional surface mesh optimization and centerline extraction of vasculatures for endovascular intervention simulation

Understanding vasculatures is important for endovascular intervention simulation (EIS). A recent trend attempts to represent vasculatures with three-dimensional (3D) surface meshes by multiview image and graphics-based techniques used in optical and laser scanners. Conversion from image volume data to 3D surface meshes, however, suffers from staircases and noise. Most existing methods in geometric processing often consider mesh smoothing in their independent settings. However, this approach either is ineffective in removing the artifacts or introduces additional types of artifacts (e.g., volume shrinkage and shape distortion). We propose an effective yet vascular-oriented mesh optimization framework. We first employ a weight-adaptive vertex resampling to eliminate staircases; then, we subdivide vascular meshes to denser ones; and finally, we exploit the constrained biquadratic Bezier surface fitting to produce optimized meshes. In EIS, clinicians are interested in the centerline of surface meshes as well as surface meshes. However, centerline extraction from vascular meshes is still challenging. We adopt an existing centerline extraction approach for vascular meshes based on the observation that human blood vessels are generally composed of piecewise cylindrical shapes. Tests on vascular data illustrate that our optimization approach achieves higher quality results regarding surface smoothness compared to previous approaches. In addition, our extraction approach effectively obtains complete centerlines from real data.

Web-Based Application for Biomedical Image Registry, Analysis, and Translation (BiRAT).

Imaging has become an invaluable tool in preclinical research for its capability to non-invasively detect and monitor disease and assess treatment response. With the increased use of preclinical imaging, large volumes of image data are being generated requiring critical data management tools. Due to proprietary issues and continuous technology development, preclinical images, unlike DICOM-based images, are often stored in an unstructured data file in company-specific proprietary formats. This limits the available DICOM-based image management database to be effectively used for preclinical applications. A centralized image registry and management tool is essential for advances in preclinical imaging research. Specifically, such tools may have a high impact in generating large image datasets for the evolving artificial intelligence applications and performing retrospective analyses of previously acquired images. In this study, a web-based server application is developed to address some of these issues. The application is designed to reflect the actual experimentation workflow maintaining detailed records of both individual images and experimental data relevant to specific studies and/or projects. The application also includes a web-based 3D/4D image viewer to easily and quickly view and evaluate images. This paper briefly describes the initial implementation of the web-based application.

Automatic Detection and Segmentation of Thrombi in Abdominal Aortic Aneurysms Using a Mask Region-Based Convolutional Neural Network with Optimized Loss Functions.

The detection and segmentation of thrombi are essential for monitoring the disease progression of abdominal aortic aneurysms (AAAs) and for patient care and management. As they have inherent capabilities to learn complex features, deep convolutional neural networks (CNNs) have been recently introduced to improve thrombus detection and segmentation. However, investigations into the use of CNN methods is in the early stages and most of the existing methods are heavily concerned with the segmentation of thrombi, which only works after they have been detected. In this work, we propose a fully automated method for the whole process of the detection and segmentation of thrombi, which is based on a well-established mask region-based convolutional neural network (Mask R-CNN) framework that we improve with optimized loss functions. The combined use of complete intersection over union (CIoU) and smooth L1 loss was designed for accurate thrombus detection and then thrombus segmentation was improved with a modified focal loss. We evaluated our method against 60 clinically approved patient studies (i.e., computed tomography angiography (CTA) image volume data) by conducting 4-fold cross-validation. The results of comparisons to multiple other state-of-the-art methods suggested the superior performance of our method, which achieved the highest F1 score for thrombus detection (0.9197) and outperformed most metrics for thrombus segmentation.

Stabilized Detection Accuracy Maximization Using Adaptive SAR Image Processing in LEO Networks

The use of low Earth orbit (LEO) satellites for world-wide surveillance services is currently actively discussed and developed because the constellation of satellites is one major approach which can provide global seamless network services. Because synthetic aperture radar (SAR), which is used for satellite image acquisition and its related signal processing, is dealing with large volumes of image data, corresponding on-demand adaptive methods for SAR image processing are essentially required for stabilized surveillance services under the consideration of data burst situations. Thus, an adaptive vision algorithm for ship detection which is one of major tasks in SAR image processing researches is proposed based on Lyapunov optimization framework, which maximizes the detection performance while satisfying stability conditions. The high-performance filters are utilized for precisely recognizing the targets whereas they introduce relatively larger delays (i.e., tradeoff exists between performances and delays). Therefore, the proposed Lyapunov optimization-based adaptive filter selection algorithm is designed based on the characteristics. Our data-intensive performance evaluation results prove that the proposed algorithm achieves desired performance improvements.

MAPPER: An Open-Source, High-Dimensional Image Analysis Pipeline Unmasks Differential Regulation of Drosophila Wing Features.

Phenomics requires quantification of large volumes of image data, necessitating high throughput image processing approaches. Existing image processing pipelines for Drosophila wings, a powerful genetic model for studying the underlying genetics for a broad range of cellular and developmental processes, are limited in speed, precision, and functional versatility. To expand on the utility of the wing as a phenotypic screening system, we developed MAPPER, an automated machine learning-based pipeline that quantifies high-dimensional phenotypic signatures, with each dimension quantifying a unique morphological feature of the Drosophila wing. MAPPER magnifies the power of Drosophila phenomics by rapidly quantifying subtle phenotypic differences in sample populations. We benchmarked MAPPER’s accuracy and precision in replicating manual measurements to demonstrate its widespread utility. The morphological features extracted using MAPPER reveal variable sexual dimorphism across Drosophila species and unique underlying sex-specific differences in morphogen signaling in male and female wings. Moreover, the length of the proximal-distal axis across the species and sexes shows a conserved scaling relationship with respect to the wing size. In sum, MAPPER is an open-source tool for rapid, high-dimensional analysis of large imaging datasets. These high-content phenomic capabilities enable rigorous and systematic identification of genotype-to-phenotype relationships in a broad range of screening and drug testing applications and amplify the potential power of multimodal genomic approaches.

Effect of Label Noise on Multi-Class Semantic Segmentation: A Case Study on Bangladesh Marine Region

ABSTRACT The volume and availability of satellite image data has greatly increased over the past few years. But, during the transmission and acquisition of these digital images, noise becomes a prevailing term. When preprocessing the data for computer vision tasks, human experts often produce noise in the labels which can downturn the performance of learning algorithms drastically. This study is directed toward finding the effect of label noise in the performance of a semantic segmentation model, namely U-net. We collected satellite images of the Bangladesh marine region for four different time frames, created patches and segmented the sediment load into five different classes. The U-Net model trained with Dec-2019 dataset yielded the best performance and we tested this model under three types of label noise (NCAR – noise completely at random, NAR – noise at random and NNAR – noise not at random) while varying their intensity gradually from low to high. The performance of the model decreased slightly as the percentage of NCAR noise is increased. NAR is found to be defiant until 20◦ of rotation, and for NNAR, the model fails to classify pixels to its correct label for maximum cases.

Reduction of satellite images size in 5G networks using machine learning algorithms

The high data volume of multispectral satellite images is compressed for better visual perception without loss of image and statistical properties of the local or global image to provide superior information for obtaining effective results using 5G networks. This compression is a technique applied to remote sensing applications to analyse the data for prediction or forecasting the real-time applications by remote sensing applications like IoT and data transmission over 5G wireless networks. The extensive data images have multiple bands, which contain earth surface/object information with various frequencies. It is difficult to handle this extensive data for processing data. The compression is mandatory to avoid this complexity by removing redundancy data, unnecessary pixel information and non-visual redundancy data between bands. There are various standard compression techniques are available like JPEG 2000, Wavelet and DCT methods. The proposed method is implemented with a combination of intra coding and machine learning algorithm. The standard compression technique does not give better results due to degradation of pixels, lack of spatial and spectral information. This paper enriches progressive results by reduced satellite images for transmission of data in IoT and 5G wireless networks, in which qualitative results are compared by standard compression technique with suitable parameters.

Use of artificial intelligence for real-time automatic quantification of left ventricular ejection fraction by a novel handheld ultrasound device

Abstract Funding Acknowledgements Type of funding sources: Other. Main funding source(s): Hellenic Society of Cardiology Background/Introduction: Transthoracic echocardiography is the most widely used imaging modality for the assessment of left ventricular ejection fraction (LVEF), which is crucial for patient care and management. Most recently, artificial intelligence (AI) has been employed to automatically identify LV endocardial boundaries and calculate LVEF. Technological advancements have enabled the development of small handheld ultrasound devices (HUDs) that can provide echocardiographic images at the point of care; however until now the HUDs have presented limited quantification capabilities. The application of AI algorithms to point-of-care echocardiography may facilitate the acquisition and accurate interpretation of a high volume of imaging data in real-time. Purpose To evaluate the accuracy of a novel HUD with AI-assisted algorithm (autoEF) to automatically calculate LVEF. Methods We prospectively included consecutive patients who were referred to a tertiary hospital echocardiography laboratory for a standard echocardiography examination. The modified biplane Simpson’s method was used to determine LV volumes and function from the apical four-chamber and apical two-chamber views. All patients were subsequently scanned with the Kosmos HUD and a fully automated estimation of the LVEF was possible after acquisition of the same two apical views within seconds by the device itself with the use of the AI-assisted autoEF algorithm. The image quality for each examination was assessed and classified as good, moderate and poor. The autoEF measurements were compared head-to-head with the manually traced biplane Simpson’s rule measurements on cart-based systems as the reference standard, using linear regression and Bland-Altman analysis. Results The study comprised 100 consecutive patients (57 ± 15 years old, 61% male), including 38 patients with abnormal LVEF &amp;lt; 50%. The image quality for the cart-based systems acquisition was assessed as good in 45%, moderate in 50% and poor in 5% of patients and for the HUD acquisition as good in 31%, moderate in 57% and poor in 12% of patients. AutoEF measurement by AI-assisted algorithm was feasible in all patients with LVEF measurements on cart-based systems. There was good agreement between the reference standard and Kosmos autoEF algorithm with a correlation coefficient r = 0.87, 95%CI 0.81-0.91, (p &amp;lt; 0.0001). The corresponding Bland-Altman plot showed a small non-significant bias of -1.42% (p = 0.058), with limits of agreement ±14.5% for the auto-EF. The paired comparison of the LVEF calculation by the 2 methods did not reveal a significant difference between reference standard and HUD autoEF [56% (IQR 40% - 62%) vs 53% (IQR 43% - 59%) respectively, p = 0.106]. Conclusion The AI-assisted autoEF algorithm in a novel HUD can accurately calculate LVEF in real-time as compared to the recommended manual biplane Simpson’s method on cart-based systems in an "all-comers" patient population. Abstract Figure. Correlation plot for LVEF measurements Abstract Figure. Bland-Altman plot

Modeling of bodies with spherical pores by generalized linear interpolation

The article offers a description of parametric objects with spherical pores by generalized linear interpolation. Increasing the volume of high-resolution image data requires the development of algorithms capable of processing large images with reduced computational costs. Numerical data on the geometry of the pores of the object under study are transformed into the geometry of bodies consisting of octagonal portions of cubic shape. Parametric porous objects can model both the shape and the isoparametric interior. Often, this type of parametric bodies is used as initial or boundary conditions in numerical modeling to demonstrate internal modeling. To form a body of complex shape, parametric solid-state elements can be connected together. The continuity between the elements can be determined in the same way as when modeling cubic parametric splines. A lot of research is devoted to the reconstruction of the geometric structure of porous materials based on digital images of objects for a better understanding and representation of physical processes in a porous medium. A detailed understanding of the microstructure can be used to determine physical properties, and then to evaluate and improve the characteristics of simulated objects and processes in them. The article presents the results of the proposed algorithm in the MathCAD environment and software processing of a porous body based on digital images.

First Gradually, Then Suddenly: Understanding the Impact of Image Compression on Object Detection Using Deep Learning.

Video surveillance systems process high volumes of image data. To enable long-term retention of recorded images and because of the data transfer limitations in geographically distributed systems, lossy compression is commonly applied to images prior to processing, but this causes a deterioration in image quality due to the removal of potentially important image details. In this paper, we investigate the impact of image compression on the performance of object detection methods based on convolutional neural networks. We focus on Joint Photographic Expert Group (JPEG) compression and thoroughly analyze a range of the performance metrics. Our experimental study, performed over a widely used object detection benchmark, assessed the robustness of nine popular object-detection deep models against varying compression characteristics. We show that our methodology can allow practitioners to establish an acceptable compression level for specific use cases; hence, it can play a key role in applications that process and store very large image data.

Face Encryption based on Feature Extraction Supported by Canny Edge Detector

Individual privacy protection is regarding a subject in surveillance and security areas. Protection of people's faces in image in linking with complex info, abused cases, involvement, and others on public broadcasting media and social networks are very important. Partial encryption human face is selected to reduce the computational requirements for huge volumes of image data and reducing the computational processing time required to conduct the entire image. In this paper, a new fast algorithm is proposed to encrypt partially human face, at first human face is detected using vole and jones algorithm, features points are extracted from face region to be encrypted, these features points are determined by using Canny edge detector since Canny edge algorithm considers the best edge detector with decent localization properties marked edges areas similar to the edges in the actual image as possible, i.e. that it is an essential feature. Furthermore, the importance of the key for the encryption process and its significant role in increasing and improving security, a new method for generating key is proposed making them resistant to attackers. The suggested encryption algorithms produced good results in the encryption process, since it has a shorter encrypting period and a greater encrypting effect, according to experimental results.

RNS-Based FPGA Accelerators for High-Quality 3D Medical Image Wavelet Processing Using Scaled Filter Coefficients

Medical imaging using different modalities has many problems. The main ones are low informativeness, various distortion noises, and a large amount of information. Fusion, denoising, and visual data compression are used to solve them in practice. Discrete wavelet transform is one way to implement various fusion, denoising, and compression methods for 2D and 3D medical image processing. Medical imaging systems produce increasingly accurate images with scanning technology and digital devices development. These images have improved quality using both higher spatial resolutions and color bit-depth. Processing a large volume of medical imaging data requires considerable resources and processing time. Modern wavelet-based devices for medical image processing do not meet the current performance demand. Hardware accelerators are being designed to solve this problem. This paper proposes new (field-programmable gate array) FPGA accelerators using wavelet processing (WP) with scaled filter coefficients (SFC) and parallel computing in residue number system (RNS) to improve the performance of high-quality 3D medical image WP systems. The computational complexity is reduced using the developed WP method with SFC and the proposed wavelet filter coefficients scaling algorithm. Parallel computing is organized in RNS using moduli sets of a particular type. Hardware implementation of 3D medical image WP using the proposed FPGA accelerators increases device performance by 2.89-3.59 times, increasing the hardware resources by 1.18-3.29 times compared to state-of-the-art solutions. The device performance improvement is achieved while maintaining high-quality 3D medical image processing in peak signal-to-noise ratio terms.