Individual abnormal behaviors vary depending on crowd sizes, contexts, and scenes. Challenges such as partial occlusions, blurring, a large number of behaviors, camera viewing occur in large-scale crowds when detecting, tracking, recognizing individuals with abnormalities. In this paper, our contribution is two-fold. First, we introduce an annotated labeled behavior Hajj dataset, HAJJv2. Second, propose two methods hybrid convolutional neural networks (CNNs) random forests (RFs) to detect...

Abstract The success of the Saudi Human Genome Program (SHGP), one top ten genomic programs worldwide, is highly dependent on population embracing concept participating in genetic testing. However, data sharing and artificial intelligence (AI) genomics are critical public issues medical care scientific research. present study was aimed to examine awareness, knowledge, attitude society towards SHGP, privacy resulting from role AI analysis regulations. Results a questionnaire survey with 804...

Breast cancer is the most common in world and second type of that causes death women. The timely accurate diagnosis breast using histopathological images crucial for patient care treatment. Pathologists can make more diagnoses with help a novel approach based on computer vision techniques. This an ensemble model two pretrained transformer models, namely, Vision Transformer (ViT) Data-Efficient Image (DeiT). ViTDeiT soft voting combines ViT DeiT model. proposed ViT-DeiT classifies...

Recognizing normality and abnormality from heart sound recordings (phonocardiograms or PCG) has promote a scientific research in cardiology. However, only small number of PCG is publicly available. Also, current recognition approaches have not reached satisfiable accuracy. In this paper, we apply transfer learning to automate the for rates normalities abnormalities. Mel Frequency Cepstrum Coefficients (MFCC) signal representation adopted transform output feature into image which used as an...

Machine Learning (ML) and Deep (DL) have been widely used in our daily lives a variety of ways with many effective stories. Also, they instrumental tackling the Coronavirus (COVID-19) epidemic, which has occurring around world. The COVID-19 epidemic caused by SARS-CoV-2 virus spread rapidly world, leading to global outbreaks. Most governments, businesses, scientific research institutions are taking part struggle stem disease. In this survey, we investigate Artificial Intelligence (AI) based...

Face detection in unconstrained environments is a challenging problem due to partial occlusions with pose variations. Existing occluded face methods require training several models, computing hand-crafted features, or both. In this paper, our contributions are two-fold. First, we propose Large-Scale Deep Learning (LSDL), method that requires single Convolutional Neural Network (CNN) model without any features detect faces. The trained large number of examples cover most and non-partial...

Abnormal behavior poses a great threat to social security and stability. The resulting violence or crime leads terrible consequences. How utilize reasonable means predict the dangerous intentions of massive crowds prevent potential hazard public is significant for security. A crowd monitoring management system an effective way detect abnormal behavior. In this article, we release unmanned aerial vehicles as well fixed ground devices achieve multi-level multi-modal behavioral sensing on...

Abstract Automatic diagnosis of malignant prostate cancer patients from mpMRI has been studied heavily in the past years. Model interpretation and domain drift have main road blocks for clinical utilization. As an extension our previous work we trained on a public cohort with 201 cropped 2.5D slices glands were used as input, optimal model searched space using autoKeras. innovative move, peripheral zone (PZ) central gland (CG) tested separately, PZ detector CG demonstrated effective...

Batch Normalization (BN) is an important preprocessing step to many deep learning applications. Since it a data-dependent process, for some homogeneous datasets redundant or even performance-degrading process. In this paper, we propose early-stage feasibility assessment method estimating the benefits of applying BN on given data batches. The proposed uses novel threshold-based approach classify training batches into two sets according their need normalization. normalization decided based...

Semantic web search engine is the new generation of conventional that brings precise and meaningful information from Internet. These engines answer user queries using Web Documents (SWDs) are found in ontologies database. It likely a query may have more than one range domain. The semantic such as Hakia, Swoogle, Watson do not identify domains ranges user's while retrieving results. Hence, retrieved results single domain query. In this paper, novel Domain Range Identifier (DRI) module...

Nucleus is the main component in a human cell. Excellent nucleus shape observation methods help endorse scientific research medical and biological fields. are performed manually by humans through labs. Abnormal (Micronucleus) caused drugs other toxical factors. Current only confined detecting segmenting nucleuses images from different tissues. None of these has tackled problem automating micronucleus recognition. In this paper, first, we apply deep transfer learning to automate recognition...

Many biometrics and security systems use facial information to obtain an individual identification recognition. Classifying a race from face image can provide strong hint search for identity criminal identification. Current classification methods are confined only constrained non-partially occluded frontal faces. Challenges remain under unconstrained environments such as partial occlusions pose variations. In this paper, we propose Convolutional Neural Network (CNN) model classify races with...

Individual abnormal behaviors vary depending on crowd sizes, contexts, and scenes. Challenges such as partial occlusions, blurring, large-number behavior, camera viewing occur in large-scale crowds when detecting, tracking, recognizing individuals with behaviors. In this paper, our contribution is twofold. First, we introduce an annotated labeled Hajj dataset (HAJJv2). Second, propose two methods of hybrid Convolutional Neural Networks (CNNs) Random Forests (RFs) to detect recognize...

Generative Adversarial Network (GAN) has made a breakthrough and great success in many research areas computer vision. Different GANs generate different outputs. In this work, we apply to handwritten Arabic characters. A basic GAN, Vanilla Deep Convolutional GAN (DCGAN), Bidirectional (BiGAN), Wasserstein (WGAN) are used. Then, the results of generated images evaluated using native-Arabic human Fréchet Inception Distance (FID). The qualitative quantitative provided for generation evaluation....

<h3>Objectives:</h3> To evaluate early performance indicators for breast cancer screening at the King Abdulaziz University Hospital in Saudi Arabia. <h3>Methods:</h3> This study retrospectively evaluated data from women who underwent their first program Jeddah, Arabia between 2012 and 2019. Data on results were used to estimate generate descriptive statistics. <h3>Results:</h3> Of 16000 invited 2019, a total of 1911 (11.9%) participated. The majority (68.8%) 40 55 years old. Based process...

Pneumonia ranks among the most prevalent lung diseases and poses a significant concern since it is one of that may lead to death around world. Diagnosing pneumonia necessitates chest X-ray substantial expertise ensure accurate assessments. Despite critical role lateral X-rays in providing additional diagnostic information alongside frontal X-rays, they have not been widely used. Obtaining from multiple perspectives crucial, significantly improving precision disease diagnosis. In this paper,...

Face detection methods have relied on face datasets for training. However, existing tend to be in small scales learning both constrained and unconstrained environments. In this paper, we first introduce our large-scale image datasets, Large-scale Labeled (LSLF) noisy Non-face (LSLNF). Our LSLF dataset consists of a large number multi-view partially occluded faces. The faces many variations color grayscale, quality, resolution, illumination, background, illusion, human face, cartoon facial...