Computer-aided detection of malignant breast tumors in ultrasound images has been receiving growing attention. In this paper, we propose a deep learning methodology to tackle problem. The training data, which contains several hundred benign and cases, was used train convolutional neural network (CNN). Three approaches are proposed: baseline approach where the CNN architecture is trained from scratch, transfer-learning pre-trained VGG16 further with images, fine-tuned parameters overcome...

Normalized cross-correlation (NCC) is used in many machine vision applications for industrial inspection. However, the high computational cost of NCC impedes real-time In this paper, we propose a modified low-complexity scheme to discover location missing integrated circuits (ICs) automatic printed-circuit board (PCB) proposed scheme, 2-D sub-images matching process are converted into 1-D feature descriptors. The running scanned vertically and horizontally, augmented with spatial statistical...

Digital pathology has recently been revolutionized by advancements in artificial intelligence, deep learning, and high-performance computing. With its advanced tools, digital can help improve speed up the diagnostic process, reduce human errors, streamline reporting step. In this paper, we report a new large red blood cell (RBC) image dataset propose two-stage learning framework for RBC segmentation classification. The is highly diverse of more than 100K RBCs containing eight different...

Mental disorders, especially schizophrenia, still pose a great challenge for diagnosis in early stages. Recently, computer-aided techniques based on resting-state functional magnetic resonance imaging (Rs-fMRI) have been developed to tackle this challenge. In work, we investigate different decision-level and feature-level fusion schemes discriminating between schizophrenic normal subjects. Four types of fMRI features are investigated, namely the regional homogeneity, voxel-mirrored homotopic...

Abstract Polymer composites have been widely exploited in numerous industries. Micro-particle fillers are typically added as reinforcement materials to improve the characteristics of these composites. In this work, organic nanoparticles date seeds were a filler for polyethylene terephthalate (PET) produce enhanced polymer nanocomposites. A date-seed nanofiller (DSN) was prepared and examined with x-ray diffraction measurements, then PET by hot compression. The PET-DSN composite...

Outlier detection (OD) is a key problem, for which numerous solutions have been proposed. To deal with the difficulties associated outlier across various domains and data characteristics, ensembles of detectors recently employed to improve performance individual detectors. In this paper, we follow an ensemble approach in good are selected through enhanced clustering-based dynamic selection (CBDS) method. method, bisecting K-means clustering algorithm partition input into clusters where every...

This paper proposes a novel approach for sparse coding that further improves upon the representation-based classification (SRC) framework. The proposed framework, Affine-Constrained Group Sparse Coding (ACGSC), extends current SRC framework to problems with multiple input samples. Geometrically, affineconstrained group essentially searches vector in convex hull spanned by vectors can best be coded using given dictionary. resulting objective function is still and efficiently optimized...

Inadequate design of emergency departments (EDs) is a major cause crowding, increased length stay, and higher mortality. The main reason behind this inadequacy the lack stakeholders’ involvement in process. This work reports analyzes results large survey requirements ED stakeholders. It then compares these with existing designs on one hand international standards other. Further, we propose new hybrid which combines both stakeholders using quality function deployment (QFD), also known as...

Analysis and classification of electromyography (EMG) signals are crucial for rehabilitation motor control. This study investigates electromyogram time-frequency representations then creates conventional deep learning models EMG signal classification. Firstly, a dataset single-channel surface has been recorded four subjects to differentiate between forearm flexion extension. Then, different have used build We compared the performance pre-trained convolutional neural network models, namely...

Recent approaches in image super-resolution suggest learning dictionary pairs to model the relationship between low-resolution and high-resolution patches with sparsity constraints on patch representation. Most of previous this direction assume for simplicity that sparse codes a are equal those corresponding patch. However, invariance assumption is not quite accurate especially large scaling factors where optimal weights indices representative features fixed across transformation. In paper,...