Search

This research explores neural network models' performance and adaptability in the context of the colorectal histology dataset as it pertains to the categorization of textures. Inception, VGG19, and MobileNet, together with their federated variations, are among the models being examined. The study includes a detailed evaluation, parameter analysis, and training information. VGG19 stands out as a particularly noteworthy high performance, with remarkable accuracy, precision, and recall. Due to its lightweight design, MobileNet performs less well, but its potential is enhanced by the addition of federated learning. The accuracy and precision of federated versions of Inception, VGG19, MobileNet, and a Lightweight MobileNet model are competitive, with FL-Lightweight MobileNet achieving outstanding results. The work has important ramifications for the field of medical image analysis since it shows how federated learning may balance the need for data confidentiality and privacy with model performance. This study marks a turning point in the development of medical imaging by opening the door to in-depth investigation into the complex interactions across federated paradigms. Furthermore, these results provide a compelling story in the wider discussion of how cutting-edge technologies and the pressing needs of contemporary healthcare might work together. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Chongqing University of Posts & Telecommunications (Natural Science Edition) is the property of Chongqing University of Posts & Telecommunications and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Jordanian Journal of Computers & Information Technology is the property of Jordanian Journal of Computers & Information Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Early detection of eye diseases is the only solution to receive timely treatment and prevent blindness. Colour fundus photography (CFP) is an effective fundus examination technique. Because of the similarity in the symptoms of eye diseases in the early stages and the difficulty in distinguishing between the type of disease, there is a need for computer-assisted automated diagnostic techniques. This study focuses on classifying an eye disease dataset using hybrid techniques based on feature extraction with fusion methods. Three strategies were designed to classify CFP images for the diagnosis of eye disease. The first method is to classify an eye disease dataset using an Artificial Neural Network (ANN) with features from the MobileNet and DenseNet121 models separately after reducing the high dimensionality and repetitive features using Principal Component Analysis (PCA). The second method is to classify the eye disease dataset using an ANN on the basis of fused features from the MobileNet and DenseNet121 models before and after reducing features. The third method is to classify the eye disease dataset using ANN based on the fused features from the MobileNet and DenseNet121 models separately with handcrafted features. Based on the fused MobileNet and handcrafted features, the ANN attained an AUC of 99.23%, an accuracy of 98.5%, a precision of 98.45%, a specificity of 99.4%, and a sensitivity of 98.75%. [ABSTRACT FROM AUTHOR]

COVID-19 is a virus that causes upper respiratory tract and lung infections. The number of cases and deaths increased daily during the pandemic. Once it is vital to diagnose such a disease in a timely manner, the researchers have focused on computer-aided diagnosis systems. Chest X-rays have helped monitor various lung diseases consisting COVID-19. In this study, we proposed a deep transfer learning approach with novel fine-tuning mechanisms to classify COVID-19 from chest X-ray images. We presented one classical and two new fine-tuning mechanisms to increase the model's performance. Two publicly available databases were combined and used for the study, which included 3616 COVID-19 and 1576 normal (healthy) and 4265 pneumonia X-ray images. The models achieved average accuracy rates of 95.62%, 96.10%, and 97.61%, respectively, for 3-class cases with fivefold cross-validation. Numerical results show that the third model reduced 81.92% of the total fine-tuning operations and achieved better results. The proposed approach is quite efficient compared with other state-of-the-art methods of detecting COVID-19. [ABSTRACT FROM AUTHOR]

Skin disease is currently considered to be one of the most common diseases in the globe. Most of the human population has experienced it at some point but not all skin illnesses are as severe as others. There are some diseases that are symptomless or show fewer symptoms. Skin cancer is a potentially fatal outcome of serious skin illnesses that might develop if they are not detected in time. Due to the fact that medical professionals aren't always quick or reliable enough to make a proper diagnosis. There is a hefty price tag attached to employing sophisticated equipment. Therefore, we propose a system capable of classifying skin diseases using deep learning approaches, such as CNN architecture and six preset models including MobileNet, VGG19, ResNet, EfficientNet, Inception, and DenseNet. Acne, blisters, cold sores, psoriasis, and vitiligo are some of the most often seen skin conditions, thus we scoured the web resources for relevant photographs of these conditions. We have applied data augmentation methods to extend the size of the dataset and include more image variations. In the validation dataset, we achieved an accuracy rate of approx 99 percent, while in the test dataset; we achieved an accuracy rate of approx 90 percent. Our proposed method would help to diagnose skin diseases in a faster and more cost-effective way. [ABSTRACT FROM AUTHOR]

Copyright of Techno.com is the property of Universitas Dian Nuswantoro, Fakultas Ilmu Komputer and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Hypertensive retinopathy (HR) is a serious eye disease that causes the retinal arteries to change. This change is mainly due to the fact of high blood pressure. Cotton wool patches, bleeding in the retina, and retinal artery constriction are affected lesions of HR symptoms. An ophthalmologist often makes the diagnosis of eye-related diseases by analyzing fundus images to identify the stages and symptoms of HR. The likelihood of vision loss can significantly decrease the initial detection of HR. In the past, a few computer-aided diagnostics (CADx) systems were developed to automatically detect HR eye-related diseases using machine learning (ML) and deep learning (DL) techniques. Compared to ML methods, the CADx systems use DL techniques that require the setting of hyperparameters, domain expert knowledge, a huge training dataset, and a high learning rate. Those CADx systems have shown to be good for automating the extraction of complex features, but they cause problems with class imbalance and overfitting. By ignoring the issues of a small dataset of HR, a high level of computational complexity, and the lack of lightweight feature descriptors, state-of-the-art efforts depend on performance enhancement. In this study, a pretrained transfer learning (TL)-based MobileNet architecture is developed by integrating dense blocks to optimize the network for the diagnosis of HR eye-related disease. We developed a lightweight HR-related eye disease diagnosis system, known as Mobile-HR, by integrating a pretrained model and dense blocks. To increase the size of the training and test datasets, we applied a data augmentation technique. The outcomes of the experiments show that the suggested approach was outperformed in many cases. This Mobile-HR system achieved an accuracy of 99% and an F1 score of 0.99 on different datasets. The results were verified by an expert ophthalmologist. These results indicate that the Mobile-HR CADx model produces positive outcomes and outperforms state-of-the-art HR systems in terms of accuracy. [ABSTRACT FROM AUTHOR]

Technological development is an inherent feature of this time, that reliance on electronic applications in all daily transactions (business management, banking, financial transfers, health, and other important aspects of life). Identifying and confirming identity is one of the complex challenges. Therefore, relying on biological properties gives reliable results. People can be identified in pictures, films, or real-time using facial recognition technology. A face individual is a unique identifying biological characteristic to authenticate them and prevents permits another person to assume that individual's identity without their knowledge or consent. This article proposes the identification model by facial individual characteristics, based on the deep neural network (DNN). The proposed method extracts the spatial information available in an image, analysis this information to extract the salient features, and makes the identifying decision based on these features. This model presents successful and promising results, the accuracy achieves by the proposed system reaches 99.5% (+/- 0.16%) and the values of the loss function reach 0.0308 over the Pins Face Recognition dataset to identify 105 subjects. [ABSTRACT FROM AUTHOR]

In recent years,with the development of unmanned driving,intelligent UAV and mobile Internet,the demand for lightweight neural network from low-power,low-cost mobile and embedded platforms is increasingly urgent.Based on the idea of deformable convolution and depthwise separable convolution,this paper presents a low-cost deformable convolution,which has the advantages of high-efficiency feature extraction ability of deformable convolution and low computational complexity of depthwise separable convolution.In addition,on the basis of applying low-cost deformable convolution and combining with the method of model structure compression,4 lightweight methods of MobileNet network are designed.Experiments on Caltech256,CIFAR100 and CIFAR10 datasets demonstrate that low-cost deformable convolution can effectively improve the classification accuracy of lightweight networks without significant increase in computational effort.Besides,the accuracy of the MobileNet network can be improved by 0.4%~1% by combining the 4 MobileNet re-lightening methods in this paper,while the network computing load can be reduced by 5% ~ 15%,which significantly improves the performance of the lightweight network and better meets the practical needs of low power consumption and low computing power.It has very important practical significance for the advancement of intelligence in the field of mobile and embedded platforms.

Environmental scientists often use birds to understand ecosystems because they are sensitive to environmental changes, but few experts are available. To make it easier to recognize bird species, an automatic system that can classify bird species is needed. There are lots of models to choose from, but some models require very high computational data when training data, reducing training time can result in less wasted electrical energy so that it can have a good effect on the environment. For this reason, it is necessary to test a model that has a small complexity in training time, whether it can produce good performance. Based on the number of neural network models available, this study will classify using the EfficientNet, EfficientNetV2, MobileNet, MobileNetV2, and NasnetMobile models to determine whether these models can have good performance. From the research results, all the models tested have good performance with an accuracy between 95% - 97%. The MobileNetV2 model has the less efficiency with the smallest training time while maintaining good performance.

With the development of deep learning and artificial intelligence, the application of lip recognition is in high demand in computer vision and human-machine interaction. Especially, utilizing automatic lip recognition technology to improve performance during social interactions for those hard of hearing, and pronunciation is one of the most promising applications of artificial intelligence in medical healthcare and rehabilitation. Lip recognition means to recognize the content expressed by the speaker by analyzing dynamic motions. Presently, lip recognition research mainly focuses on the algorithms and computational performance, but there are relatively few research articles on its practical application. In order to amend that, this paper focuses on the research of a deep learning-based lip recognition application system, i.e., the design and development of a speech correction system for the hearing impaired, which aims to lay the foundation for the comprehensive implementation of automatic lip recognition technology in the future. First, we used a MobileNet lightweight network to extract spatial features from the original lip image; the extracted features are robust and fault-tolerant. Then, the gated recurrent unit (GRU) network was used to further extract the 2D image features and temporal features of the lip. To further improve the recognition rate, based on the GRU network, we incorporated an attention mechanism; the performance of this model is illustrated through a large number of experiments. Meanwhile, we constructed a lip similarity matching system to assist hearing-impaired people in learning and correcting their mouth shape with correct pronunciation. The experiments finally show that this system is highly feasible and effective.

Monkeypox is a viral disease that has recently rapidly spread. Experts have trouble diagnosing the disease because it is similar to other smallpox diseases. For this reason, researchers are working on artificial intelligence-based computer vision systems for the diagnosis of monkeypox to make it easier for experts, but a professional dataset has not yet been created. Instead, studies have been carried out on datasets obtained by collecting informal images from the Internet. The accuracy of state-of-the-art deep learning models on these datasets is unknown. Therefore, in this study, monkeypox disease was detected in cowpox, smallpox, and chickenpox diseases using the pre-trained deep learning models VGG-19, VGG-16, MobileNet V2, GoogLeNet, and EfficientNet-B0. In experimental studies on the original and augmented datasets, MobileNet V2 achieved the highest classification accuracy of 99.25% on the augmented dataset. In contrast, the VGG-19 model achieved the highest classification accuracy with 78.82% of the original data. Considering these results, the shallow model yielded better results for the datasets with fewer images. When the amount of data increased, the success of deep networks was better because the weights of the deep models were updated at the desired level. [ABSTRACT FROM AUTHOR]

Simple Summary: The histopathological detection of these malignancies is a vital element in determining the optimal solution. Timely and initial diagnosis of the sickness on either front diminishes the possibility of death. Deep learning (DL) and machine learning (ML) methods are used to hasten such cancer recognition, allowing the research community to examine more patients in a much shorter period and at a less cost. Cancer is a deadly disease caused by various biochemical abnormalities and genetic diseases. Colon and lung cancer have developed as two major causes of disability and death in human beings. The histopathological detection of these malignancies is a vital element in determining the optimal solution. Timely and initial diagnosis of the sickness on either front diminishes the possibility of death. Deep learning (DL) and machine learning (ML) methods are used to hasten such cancer recognition, allowing the research community to examine more patients in a much shorter period and at a less cost. This study introduces a marine predator's algorithm with deep learning as a lung and colon cancer classification (MPADL-LC3) technique. The presented MPADL-LC3 technique aims to properly discriminate different types of lung and colon cancer on histopathological images. To accomplish this, the MPADL-LC3 technique employs CLAHE-based contrast enhancement as a pre-processing step. In addition, the MPADL-LC3 technique applies MobileNet to derive feature vector generation. Meanwhile, the MPADL-LC3 technique employs MPA as a hyperparameter optimizer. Furthermore, deep belief networks (DBN) can be applied for lung and color classification. The simulation values of the MPADL-LC3 technique were examined on benchmark datasets. The comparison study highlighted the enhanced outcomes of the MPADL-LC3 system in terms of different measures. [ABSTRACT FROM AUTHOR]

Myocardial Infarction (MI) is a heart disease due to lack of blood and oxygen flow to some part of the heart muscle. MI is a leading cause of death globally. To decrease the likelihood of MI complications, it is essential to diagnose and treat MI promptly. A Light Weight Multi-branch Network (LW-MN) based feature extraction and classification of MI from Electrocardiogram image is proposed. Extracting the information from Electrocardiogram images without segmenting the leads separately may result in loss of information. Segmenting the 12 leads by detecting the text above each lead will extract the information accurately without any information loss. Light Weight model used to extract features and fusing all twelve leads features using the depth fusion method. The outcome from the fusion method is fed into DenseNet-161 classification algorithm. LW-MN model acquired a classification accuracy, specificity, sensitivity, and F1-score of 96.09%, 97.33%, 97.78%, and 95.18%. [ABSTRACT FROM AUTHOR]

Purpose. Research on the dependence of the amount of microcontroller memory used and the image recognition time on the type of convolutional neural network MobileNet V1 or V2 and their hyperparameters. Methodology. Creating a database of images necessary for training a neural network using the Edge Impulse software platform with subsequent loading of the network into the memory of a 32-bit microcontroller ESP32 for practical evaluation of network characteristics. Findings. A comparison of the characteristics of MobileNetV1 and MobileNetV2 neural networks was made. Experiments were conducted to determine the dependence of the recognition time of selected objects on the image, the amount of used RAM and program memory, based on the ESP-EYE microcontroller with a camera, depending on the network width ratio, image size and convolutional neural architecture network. It was determined that the time of image classification for the MobileNetV2 network model takes from three to ten seconds or more, which is not acceptable for fast recognition tasks. It was also found that the microcontroller program memory is insufficient to analyze the 160 by 160 pixel images with the maximum network width of the MobileNetV2 model. Using the MobileNetV1 network provides slightly lower recognition accuracy, but requires significantly less microcontroller resources and time. Originality. The peculiarities of the use, possibilities and limitations of neural networks for image recognition in systems based on ESP32 microcontrollers have been verified in practice. The dependence of the amount of memory used on the microcontroller and the time of image recognition on the type of convolutional neuron was established in order to select additional image processing tools to improve the quality of recognition. Practical value. The obtained results allow the proper selection of the MobileNetV1 or MobileNetV2 neural network depending on the specific tasks of image recognition by systems on microcontrollers. [ABSTRACT FROM AUTHOR]

Human-computer interaction, human-robot interaction, robotics, healthcare systems, health assistive technologies, automotive user interfaces, crisis management, disaster relief, entertainment, and contactless communication in smart devices are just a few of the practical applications for hand gesture recognition. In this work, we propose two novel machine learning models for hand gesture recognition using three publicly available datasets, NUS Hand Posture Dataset I, Turkey Ankara Ayrancı Anadolu High School's Sign Language Digits Dataset, and American Sign Language dataset. The developed models based on vision transformer and lightweight Convolutional Neural Networks are deployed on an end-to-end edge computing system that can accurately provide the classification of hand gestures. The edge computing system presented here utilizes Raspberry Pi. The designed models achieve an accuracy of 90–99% on the test datasets. The performance of the proposed models is also compared to the pre-trained models such as MobileNet, ResNet, and VGGNet. [ABSTRACT FROM AUTHOR]

Copyright of Dicle University Journal of Engineering / Dicle Üniversitesi Mühendislik Dergisi is the property of Dicle Universitesi and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Malaria is predominant in many subtropical nations with little health-monitoring infrastructure. To forecast malaria and condense the disease's impact on the population, time series prediction models are necessary. The conventional technique of detecting malaria disease is for certified technicians to examine blood smears visually for parasite-infected RBC (red blood cells) underneath a microscope. This procedure is ineffective, and the diagnosis depends on the individual performing the test and his/her experience. Automatic image identification systems based on machine learning have previously been used to diagnose malaria blood smears. However, so far, the practical performance has been insufficient. In this paper, we have made a performance analysis of deep learning algorithms in the diagnosis of malaria disease. We have used Neural Network models like CNN, MobileNetV2, and ResNet50 to perform this analysis. The dataset was extracted from the National Institutes of Health (NIH) website and consisted of 27,558 photos, including 13,780 parasitized cell images and 13,778 uninfected cell images. In conclusion, the MobileNetV2 model outperformed by achieving an accuracy rate of 97.06% for better disease detection. Also, other metrics like training and testing loss, precision, recall, fi-score, and ROC curve were calculated to validate the considered models. [ABSTRACT FROM AUTHOR]

Individuals who are deaf or those having hearing problems can communicate through alternative modes of communication, such as sign language, both within and outside their community. As a result, the sign serves as a tool that mediated communication goals. With the aid of recent advances in computer vision, there has been promising progress in the fields of motion and gesture detection using deep learning. In this work, convolutional neural networks are utilised to develop a model for recognising the Arabic language's alphabet signs in order to aid communication goal. The research used the Arabic Alphabets Sign Language Dataset (ArASL2018), which contains sets of images indicating each specific sign for each letter of the alphabet. Experimental analysis was carried out using the proposed model. The results of the analysis reveal a recognition accuracy of 94.46%. Furthermore, the comparative analysis with the previous studies indicates that the proposed model outperformed all of them in terms of recognition accuracy. [ABSTRACT FROM AUTHOR]

The current convolution neural network approaches have attracted extensive interest because the performance is better than that of conventional machine learning methods in the plant disease recognition. However, there are still facing challenges. For instance, the image background sometime is complex, and the model can detect plant lesions, but it is difficult to use and detect the specific pest position. The high complexity of the model is not conducive to the deployment and development of mobile software. Even the dataset has problems such as labeling errors and few positive or negative samples, which restrict the development of disease recognition. In this study, we investigate the deep convolution networks based on deep transfer learning for plant disease recognition. We propose a model called as Selective Kernel MobileNet (SK-MobileNet), which is lightweight enough to greatly reduce the computing cost when deployed to servers. Experimental results show that the proposed approach reaches the accuracy of 99.28% in the public dataset. The proposed approach illustrates a significant increase in the efficiency with the lower complexity compared to other existing methods. [ABSTRACT FROM AUTHOR]

Tower state recognition is of great importance in smart electric power management. Most previous efforts were devoted on the guyed towers and tower structures or materials to monitor the structural stability of towers through sensors fixed on them. Actually, identifying their states as a whole is also needed for remote management. This study addresses the task of tower state recognition in high-resolution remote-sensing images under the framework of deep learning. To achieve this goal, a new remote-sensing image dataset is constructed. Technically then, an ensemble learning strategy is developed to classify the states of the transmission towers. Specifically, MobileNetV3 is taken as the weak classifier as it is simple yet parameter-light in the family of convolutional networks. Along the line of ensemble learning with the principal of Adaboost in machine learning, an Adaboost-style network paradigm, named as Adaboost-MobileNetV3 in this paper, is finally proposed to recognize the states of towers. To evaluate its effectiveness, a series of deep convolutional networks are compared with the proposed network. The extensive comparative experiments indicate that the proposed Adaboost-MobileNetV3 network is suitable for practical tower state recognition, which achieves about 90% accuracy with at least about 4% higher than these compared models. [ABSTRACT FROM AUTHOR]