Search

The use of facial recognition method has spread to many different industries, most notably smart attendance systems. But one major obstacle these systems must overcome is spoofing, which is the practice of people trying to trick the system by utilizing counterfeit or modified photos or videos. In the context of a smart attendance framework, this paper presents a smart anti-spoofing mechanism that can distinguish amongst real and fake faces with accuracy. The system that has been developed entails the generation of datasets that contain both actual and fake facial photos and videos, in addition to the application of a feature extraction and classification pipeline that is driven by deep learning. Furthermore, a number of data augmentation techniques, including translation, scaling, and rotation, are used to improve the effectiveness of the suggested system against different spoofing strategies. Empirical results show that these strategies support increased accuracy and durability in thwarting various spoofing attack types. With a 97% recognition rate overall, our anti-spoofing/mocking technology affords a stable and secure foundation for intelligent attendance systems. The study emphasizes how effective it is to create a strong anti-spoofing system specifically for keen attendance applications by integrating deep learning and soft computing models. [ABSTRACT FROM AUTHOR]

Attendance monitoring is an essential task for any organization and traditional methods such as pen and paper or electronic card systems have their limitations with advancements. The proposed model is a real time face presence monitoring system using Raspberry Pi. The system captures image of individuals using camera and processes the images using deep learning algorithms to detect faces. These faces are then matched against a pre-existing database of known individuals. Door interlock controls add an extra layer of security, ensuring that only authorized person have access to restricted areas. The system provides a user-friendly interface to view everyone's attendance records and generate attendance management reports. The proposed system provides a reliable and accurate solutions for attendance tracking and can be used in various organizations. [ABSTRACT FROM AUTHOR]

With the rapid evolution of the Internet, the landscape of cyber-attacks is constantly changing, leading to a rather pessimistic outlook on cyber security. This article delves into the realm of network analysis for intrusion detection, specifically focusing on the implementation of Methods for Deep Learning (DL) and Machine Learning (ML). A comprehensive tutorial description is provided for each ML/DL method, along with an examination of relevant research papers. These papers were meticulously sorted, examined, and condensed according to their thermal or temporal correlations. Given the paramount importance of data in ML/DL methods, the article also sheds light on commonly utilized network datasets within this domain. Furthermore, it addresses the challenges associated with employing ML/DL for cyber security and offers valuable suggestions for future research directions. Interestingly, the KDD data collection has become a recognized industry standard for intrusion detection methods. Just as crucial as the intense efforts to enhance intrusion detection strategies is the caliber of the data utilized to train and evaluate the detection model. With a particular focus on four different attribute classes—Basic, Content, Traffic, and Host—this research conducts a thorough study of the KDD data set. To categorize these attributes, the Modified Random Forest (MRF) approach is employed. We use the Modified Random Forest (MRF) technique to classify these properties. Two important assessment metrics that are examined are the Rate of Detection (DR) and the Rate of False Alarm (FAR), both of which are essential for an efficient Intrusion Detection System (IDS). The data set's empirical analysis uncovers the unique contributions of each attribute class to DR and FAR, making it easier to optimize the data set to maximize DR and minimize FAR. [ABSTRACT FROM AUTHOR]

The task of recognizing and categorizing vehicles in videos and images as objects poses a considerable challenge in terms of appearance-based representation. However, it holds great importance in the practical implementation of Intelligent Transportation Systems (ITS), particularly in real-time applications. The fast advancement of deep learning has resulted in an increasing need within the computer vision field for the development of efficient, robust, and outstanding services across diverse domains. This paper provides an extensive analysis of various methodologies for vehicle detection and classification, along with their utilization in real-time targets, estimating the density of traffic, and related domains through the implementation of Deep Learning techniques. The major findings of our survey highlight crucial insights obtained from an extensive analysis of existing literature, shedding light on the current state-of-the-art techniques. Through a comprehensive review of deep learning methodologies, performance metrics, benchmark datasets, and a comprehensive exploration of the challenges encountered, our survey offers valuable contributions to the field. By synthesizing and presenting the collective knowledge in this domain, our paper serves as a key resource for researchers and practitioners alike, providing a holistic understanding of the advancements and challenges in vehicle identification and categorization within deep learning architectures. [ABSTRACT FROM AUTHOR]

The accurate staging of bladder cancer is paramount for guiding initial treatment strategies, with non-muscle invasive bladder cancer often necessitating transurethral resection and muscle-invasive bladder cancer requiring adjuvant chemotherapy followed by cystectomy. This paper presents a sophisticated deep-learning model designed to enhance accuracy in bladder cancer staging, thereby facilitating precise clinical management and ultimately improving patient outcomes. Utilizing a comprehensive review of recent advancements, this paper provides an overview of the application of deep learning techniques in the imaging assessment of bladder cancer. It examines current methodologies employed in diagnosing, staging, and managing the treatment of bladder cancer through the analysis of medical images. Additionally, the paper discusses the existing constraints and challenges faced in the implementation of deep learning. The methodology employed in this study involves a systematic review of recent literature on deep learning applications in bladder cancer imaging analysis. Various deep-learning architectures are examined for their efficacy in diagnosing and staging bladder cancer. The review encompasses studies that utilize diverse imaging modalities, such as X-ray, CT scans, histology slides, and cystoscopy images, to provide a comprehensive understanding of the current landscape of deep learning-based approaches in bladder cancer assessment. The study conducts a comparative analysis of different deep-learning methodologies used in bladder cancer diagnosis and staging, evaluating their respective strengths, limitations, and performance metrics. By synthesizing findings from various studies, the paper discusses the relative efficacy of deep learning models in accurately determining the bladder cancer stage and guiding treatment decisions. Furthermore, it examines the challenges associated with data availability, model interpretability, and generalizability, offering insights into areas requiring further research and development. The results highlight the potential of deep learning in revolutionizing bladder cancer imaging assessment, with implications for enhancing clinical decision-making and improving patient outcomes. The discussion underscores the importance of addressing existing constraints and advancing the field through interdisciplinary collaboration and innovation. [ABSTRACT FROM AUTHOR]

One of the leading causes of loss of life globally is heart disease. This refers to conditions that affect the heart's normal functioning, such as heart attacks and heart failure. Heart disease stands out as a major contributor to global mortality, and early detection performs a crucial function in enhancing the chances of recovery. In trendy years, deep learning (DL) strategies have established promising outcomes in numerous medical applications, which include several types of heart illnesses. DL algorithms can robotically extract applicable capabilities from unprocessed information, making them adequately appropriate for reading complex clinical datasets. By training on big quantities of categorized information, deep studying fashions can discover ways as needed to classify remarkable varieties of heart ailments based totally on numerous enter modalities collectively with electrocardiograms (ECG's), echocardiograms, and scientific pictures. Several studies have confirmed the effectiveness of DL in classifying heart diseases with excessive accuracy and performance. These models are constrained in their ability to effectively identify precise cardiac conditions. However also anticipate future cardiovascular occasions based on risk factors. To conclude, the use of DL to know strategies for the magnificence of heart illnesses shows splendid potential in enhancing evaluation and prognosis. Further investigation and progress within this domain may lead to more accurate and efficient approaches for early detection. This paper comprehensively surveys DL strategies for detecting heart diseases. Several important points that the researchers rely on in their work to obtain the best results were emphasized. Firstly, a comprehensive comparison was made between the research papers, then a focus was made on the techniques used in feature extraction and their impact on the work of deep learning techniques while considering that the same dataset was used. Finally, it was concluded that the highest accuracy obtained when CNN and continuous wavelet transform (CWT(algorithms were applied was 99.6%. The highest accuracy reached when using an elephant herding optimizer turned restricted Boltzmann machine network (EHO-RBM) in 2023 was 99.96%. Multiple Deep Learning methodologies were addressed in this paper. The implementation of these methodologies is categorized according to distinct metrics, and the datasets used for preparation and testing undergo thorough analysis. A complete evaluation of DL strategies for heart disease was given in this overview paper. [ABSTRACT FROM AUTHOR]

The objective of speech separation is to extract the desired speech signal from its surrounding interference, traditionally conceptualised within the framework of signal processing challenges. However, a contemporary perspective redefines speech separation as a supervised learning problem, emphasizing the acquisition of discriminative patterns from training data pertaining to speech, speakers, and background noise. This paper conducts a thorough investigation of the research landscape pertaining to deep learning-based supervised speech separation in recent years. The transition towards a supervised learning framework, coupled with the integration of advanced Deep Learning (DL) methodologies, has markedly expedited progress and substantially enhanced separation performance. Through presenting a comprehensive overview, this paper seeks to distil essential insights and underscore pivotal advancements within the domain of supervised speech separation. By offering a thorough overview, the objective of this paper is to encapsulate key insights and advancements in the field of supervised speech separation, providing valuable perspectives on the current state-of-the-art in this dynamically evolving field. [ABSTRACT FROM AUTHOR]

Human pose detection plays a pivotal role in various applications, particularly in the context of exercise and fitness monitoring. This survey paper provides an in-depth analysis of the state-of-the-art techniques and advancements in human pose detection for exercise-related activities. We categorize and review key methodologies and evaluation metrics used in this domain. Beginning with an overview of the importance of human pose detection in exercise, we delve into the evolution of pose estimation methods, from traditional computer vision techniques to the rise of deep learning-based approaches. We discuss the challenges specific to exercise scenarios, such as real-time performance, multi-person tracking, and robustness in varied environments. This survey not only summarizes the current state of the art but also identifies emerging trends and future directions in human pose detection for exercise. We discuss the proposed methodology for human pose detection for exercises. By providing a comprehensive overview of the field, this survey paper aims to serve as a valuable resource for researchers, practitioners, and fitness technology developers interested in leveraging human pose detection for exercise monitoring and improvement. [ABSTRACT FROM AUTHOR]

People in Tamil Nadu have been more interested in history during the past years. When people see the old script inscribed on stone walls, potteries, copper plates, and Palm Manuscripts, history is more intriguing. But people have to learn old scripts in order to read these inscriptions, so not many people are capable of reading them. A ground breaking effort that educates the public about these ancient scripts and eventually sparks their interest in history is the use of deep learning algorithms for script recognition. Recognizing ancient Tamil characters is a difficult challenge for epigraphers because the language has evolved over the previous three thousand years with distinct characters set. Though there have been a few studies published on the recognition of ancient Tamil alphabets, there have been no papers on the recognition of ancient Tamil numbers. This project's main objective is to use convolution neural network (CNN) to recognize Tamil digits used between 9th to 12th century in Tamil inscriptions. In this paper work, the inscription photographs acquired from various Tamil temples are pre-processed. The color photographs were cropped, converted to grayscale and then to binary images during the pre-processing process. These pre-processed images are used as training and validation datasets for the model. [ABSTRACT FROM AUTHOR]

Automated vehicle number plate recognition or Number Plate Recognition (NPR) is crucial in various applications, including law enforcement, traffic management, and toll collection systems. However, the traditional method of manual inspection of images is time-consuming and prone to errors. With the advancements in computer vision and deep learning techniques, identifying vehicle number plates has become more efficient and accurate. In this paper, we present a methodology for identifying vehicle number plates using the VGG16 convolutional neural network. We collected a dataset of vehicle images with their corresponding number plates and preprocessed the data to train the CNN using algorithms like transfer learning. The VGG16 CNN has been pre-trained on a large data set that is the ImageNet dataset, which makes it an excellent candidate for fine-tuning on specific tasks like vehicle number plate recognition. After training, we evaluated the performance of the CNN on a test dataset and deployed it in a real-world application. Our results show that the VGG16 CNN can accurately recognize number plates on different types of vehicles in various lighting conditions. This methodology improves efficiency and accuracy while reducing manual inspection errors. It can be used as an essential tool for automated number plate recognition in various applications, enabling faster processing times and reducing costs associated with manual inspection. The methodology presented in this paper demonstrates the effectiveness of deep learning techniques in automating vehicle number plate recognition. The VGG16 CNN architecture, combined with transfer learning and preprocessing techniques, can accurately recognize number plates, providing a reliable and efficient solution for automated number plate recognition system. [ABSTRACT FROM AUTHOR]

Water quality prediction has a crucial significance in ensuring the safety and usability of water resources. In this paper, we propose a CNN with BiLSTM model for accurate water quality prediction. The model combines convolutional neural network layers with BiLSTM to effectively capture the spatio-temporal dependencies present in the data. Experimental results on the water quality dataset demonstrate the superior performance of the proposed model. The model achieves a validation accuracy of 92.83%. These results surpass those obtained by other models such as GRU, LSTM, and CNN with LSTM. The CNN with BiLSTM model's ability to effectively learn and classify water samples based on provided features highlights its potential for reliable water quality prediction. This paper contributes to the field of water quality prediction by providing a robust model that can aid in ensuring the safety and suitability of water resources. [ABSTRACT FROM AUTHOR]

As the tech industry continues to embrace the Internet of Things (IoT), a multitude of wireless devices are being developed to track various infrastructures. These devices generate vast amounts of statistics from domains such as medicine, supply chain, power, agricultural analytics or intelligence, BAS (including HVAC) and similar data-producing industries [1]. To effectively utilize this data and facilitate strategic decision-making, big data techniques are crucial in IoT processes. They serve as valuable instruments for real-time data visualization, enabling the extraction of useful information. This paper aims to provide an extensive analysis of the benefits of big data visualization on IoT approaches, related applications/ softwares and techniques. With the focus on visual analytics, our work situates data visualization as a part of the visual analysis phase. It is a review of the available tools for data visualization and provides applicatory guidelines for them, taking into account the specific requirements of each individual case. In spite of big data methods being applied across various IoT domains, the paper delves into the challenges of visualization and the influence of big data on shaping the IoT landscape. This article includes a studyof existing works to establish its foundation, wherein, in spite of not presenting any specific findings, it presents an overview of the progress made thus far in big data visualization along with application of deep learning in the sphere of IoT. The paper also highlights big data's prominent role in IoT visualization, with the focus on illustrating the key concepts of real-time visualizationof big data. [ABSTRACT FROM AUTHOR]

Rice is a vital staple crop, and the presence of diseases can significantly impact its production. Consequently, there is an increasing demand to create efficient strategies for recognizing and controlling diseases in rice plants. This paper provides a comprehensive review of the recent advancements in disease identification techniques for rice plants, focusing on the integration of image processing, machine learning, and deep learning methodologies. Despite the progress made, there remains a need for further research to identify and classify additional rice plant diseases beyond the major ones discussed in this paper. By continually exploring new techniques and expanding the scope of disease identification, researchers have the potential to enhance the overall effectiveness of rice disease detection systems. This review serves as a valuable resource for researchers and practitioners in the field, offering insights into the current state of the art and suggesting future directions for advancements in the identification and management of rice plant diseases introduced. [ABSTRACT FROM AUTHOR]

Various object detection models are presented in this paper. Definition of object is well defined in order to understand object detection models. Now-a-days the object detection which is computer vision and image processing related technology is widely used. Applications of object detections are infinite as it can be used all over like objects tracking, surveillance of videos, face recognition, detection of anomaly, counting of people, and the records still continue to many more. This paper compares various object detection algorithms the basic splitting is based on machine learning based and deep learning based. Fundamental objective of the paper focuseson Single stage object detectors (SSD) like You Only Look Once (YOLO) and Two stage object detectors as RCNN, Fast RCNN, etc. This paper's main goal is to highlight such object detection models and their use in various fields. [ABSTRACT FROM AUTHOR]

Deep learning-based gender detection is demonstrating its effectiveness in a variety of areas, including healthcare, marketing, security, and many others. Deep learning models have boosted their capability to identify objects in images and videos. This paper includes an extensive description of gender detection using the framework YOLOv7, including architecture, limitations, and performance. The research study analyzes YOLOv7's performance in detecting gender from facial images and videos, with a particular focus on lightning-fastness, precision, and robustness. The experimental results prove that YOLOv7 outscored YOLOv7x. The paper addresses the YOLOv7 framework's potential in various areas such as healthcare, marketing, and security. Overall, this paper demonstrates the potency of YOLOv7 in detecting gender. [ABSTRACT FROM AUTHOR]

Early identification and accurate diagnosis are crucial for successful therapy and improved patient outcomes of brain tumors. The study looks at how deep learning may help detect brain tumors, including how convolutional neural networks (CNNs) can change the face of diagnostic imaging. This underscores the need of precise and timely identification of brain cancer for efficient therapy and enhanced patient results. The study discusses the challenges associated with tumor segmentation and classification, such as variations in size, shape, and location. To overcome these challenges, deep learning techniques are utilized to analyze brain scans and identify tumor abnormalities. However, the paper underscores the importance of these models serving as tools to assist medical professionals rather than replacing their expertise. Additionally, the paper presents a comparative study that evaluates different techniques, artificial neural networks (ANN) achieve the highest accuracy at 99%, followed by Support Vector Machine (SVM) at 98.9%. vision transformer (ViT), Extreme Gradient Boosting (XG Boost), CNN-based dense Efficient Net, and Deep Neural Network (DNN) also show high accuracies above 98%. However, U-Net, Alex Net, hybrid ensemble method (Random Forest (RF), K-Nearest Neighbour (KNN), Decision Tree (DT), and Region-based convolutional neural networks (R-CNN)) have lower accuracies ranging from 91.66% to 97.305%. Overall, the findings highlight the potential of deep learning in enhancing the efficiency, precision of brain tumor diagnosis, leads to earlier treatment and enhanced patient prognosis. [ABSTRACT FROM AUTHOR]

Rice is the third highest production crop globally. There are many diseases that will affect the productivity of rice plant. The common rice plant diseases are Rice blast, Shealh Blight, Bacterial leaf blight, brown spot and Rice leaf streak. Early identification of these diseases will play crucial role in crop yield. Recent deep learning architectures can be trained in such a way that it can detect multiple diseases at once. The article aims to provide overview of all those research works that uses deep learning architectures for the identification of rice disease. This paper begins with the introduction of the importance of rice plant disease identification. Then the paper explains the various rice diseases, followed by explaining general deep learning architecture that is used for the identification of Rice plant disease. Then the paper aims to provide information about research papers that uses deep learning architectures for rice disease identification and will try to spot any limitations. At last, the paper aims to provide the future scope of the same. [ABSTRACT FROM AUTHOR]

With the increase in frequency and intensity of natural disasters induced by extreme weather events globally, there are growing threats to human life and property safety. Effective utilization of unmanned aerial vehicle (UAV) vision techniques for rapid disaster assessment and relief is of great value, but still faces many challenges. The purpose of this study is to provide a comprehensive review on the application status, technological progress, and existing problems of UAV vision technology in disaster response. Through extensive investigation of relevant literature, this paper compares and analyze the advantages and disadvantages of key UAV visual techniques for disaster response. The mainstream algorithms and future improvements of several critical technologies are summarized, including image stabilization, target recognition, and three-dimensional reconstruction. Image stabilization technology compensates for UAV motion to obtain clear and stable video, crucial for further analysis. Both mechanical and digital methods have been explored. Target recognition enables UAVs to automatically detect disaster-related entities like survivors, dangerous areas, and blocked roads. Deep learning methods are now dominant. Three-dimensional reconstruction builds precise environment models, providing essential data for assessment and planning. Multi-view methods improve accuracy. In addition, this paper discusses the trends and open challenges faced by UAV vision applications in complex disaster environments, such as robustness, real-time performance, energy efficiency, and data security. Integrating multiple visual modules into a unified system is an important direction. This study provides valuable insights to promote the engineering development and field deployment of UAV vision technology for disaster response. The analysis and summaries serve as important references for future research, development, and applications in this domain. [ABSTRACT FROM AUTHOR]

This academic paper explores the complex utilization of integrated task and motion planning (ITMP) in multi-robot systems. Given the rapid advancement of robotics and automation, multi-robot systems have gained increasing attention in various applications, such as smart manufacturing, logistics storage, environmental surveillance, and dual or multi-armed robotic platforms. The fusion of task and motion planning has become a key technology in the development of multi-robot systems. Initially, the paper suggests numerous strategies to integrate task and motion data to improve planning efficiency, thus boosting the motion planning efficiency of robots during task execution. It then investigates the use of learning techniques, such as reinforcement learning and deep learning, to enhance task and motion planning. Ultimately, it presents experimental evidence to support the claim that the integration of task and motion planning can effectively increase the efficiency of multi-robot systems and optimize their performance. The paper also examines modeling methods for multi-robot systems, typically incorporating task data with motion data to better guide robots during task execution. It presents shared space graph methods, sample-based planning methods, and market-based methods. Furthermore, it reviews learning models used to train and optimize machine learning models in ITMP. It introduces the INFORMED algorithm, a learning search issue in the field of task and motion planning that employs multilayer perceptron (MLP) and graph neural networks (GNN) to model relationships between streams and objects. Through this paper, the authors aim to offer valuable insights for the research and application of ITMP in multi-robot systems, thereby promoting further progress in this area. [ABSTRACT FROM AUTHOR]

In the present day, deep learning applications are widely used in every field of development. In this paper, such a type of application is discussed in real-time. Mura dataset is the largest abnormal dataset given by the University of Stanford. In this research paper, we will understand the pre-processing of the image dataset. Kera's and TensorFlow libraries are used to build this model. We are classifying images using different ML models where accuracy is compared to get the best model build. This paper gives us a clear understanding of building deep learning models, understanding pre-processing techniques, and working on heavy data. This research's main agenda is to create a model which can classify bone fracture images and also give a small path for further development in the deep learning area. In this project we are using Resnet50, InceptionV3, Mobile Net and DenseNet with optimizer Adam and comparing their accuracies graphically. [ABSTRACT FROM AUTHOR]

At present, detection and recognition of animals remain a complicated task and no distinctive technique can solve all situations strongly and effectively. Animal detection in images is challenging because of the complexity of wild environments. Including several different things: The large differences in shapes and color appearances of variations of objects which belong to the same class, and Lighting/Illumination Conditions. This paper presents an effective and efficient animal detection system for Birds detection in images. MobileNet SSD model is a Single Shot Multibox Detection (SSD) suggested in this work, which is capable of handling different shapes and view angles of the objects. This approach allows extracting features of image regions and using deep learning techniques for classifying regions into two categories: (bird and non-animal). The experimental results demonstrated the efficiency of classification and detection of single and multi-animals even in difficult environmental conditions. The accuracy of detection and classification is up to 98.8%. This methodology is more accurate than other similar approaches. This paper exhibits a method for training a convolutional neural network (CNN) based on object detection classifiers. [ABSTRACT FROM AUTHOR]

The rise of cyber-attacks such as Distributed Denial of Service (DDoS) and SQL injection has become a significant concern for organizations and individuals who rely on the internet. Traditional detection methods have become increasingly ineffective in addressing these challenges, necessitating the development of new and innovative solutions. This paper proposes using Convolutional Neural Networks (CNNs) to detect DDoS and SQL injection attacks. Our paper proposes a Convolutional Neural Network model with a multi-layer neural and relu activation function optimizer that reaches a higher degree of precision than previous Deep Learning models. In this study, we tested the relatively new dataset CIC-IDS-2018, which contains different types of attacks. With this dataset, our model achieves an unprecedented accuracy of>96%, minimizing computational time. [ABSTRACT FROM AUTHOR]

Monkeypox disease was declared a Public Health Emergency of International Concern (PHEIC) by the World Health Organization (WHO) on July 23rd, 2022. The purpose of this designation is to encourage the global community to take coordinated measures to contain the spread of the disease and protect communities. The recovery period for Monkeypox disease typically ranges from 2 to 4 weeks, and in some cases, the disease can resolve spontaneously with or without treatment. However, severe cases have been reported, which can lead to fatalities. The fatality ratio for Monkeypox disease is estimated between 3% to 6% at present. It is essential to take precautionary measures to prevent the spread of this disease and seek medical attention if any symptoms are observed. Monkeypox is not highly contagious, and it generally spreads through close contact with an infected person. However, in May 2022, a significant number of cases of Monkeypox disease were identified, indicating that the disease was spreading rapidly. In light of this situation, it is crucial to track and isolate suspected cases promptly to prevent further transmission. Rapid and effective contact tracing, along with timely isolation of cases, is essential to contain the spread of the disease and prevent it from becoming an epidemic. Identifying, controlling, and taking measures to control the spread of Monkeypox disease is of utmost importance. To aid in this effort, this paper aims to create a large-scale dataset of Monkeypox disease images using Generative Adversarial Networks (GANs). GANs have proven to be highly effective in generating high-quality images and can aid in creating a more comprehensive and diverse dataset. The generated images will be used to train different deep learning models for image classification to diagnose Monkeypox disease. The deep learning models will be compared, and their performance will be evaluated using images of infected human skin with Monkeypox disease and healthy skin images. The results obtained from this paper will improve the accuracy of previous works, aiding in the early diagnosis and treatment of Monkeypox disease. [ABSTRACT FROM AUTHOR]

This paper reviewed the Hand Sign Recognition system from 2018 to 2022. Based on the review, much research is done on Deep learning-based hand sign recognition. Many of these studies are published annually in article, journals, and conference proceedings. Additionally, we also examined the performance of the HGR system in terms of recognition accuracy. In this paper we concentrate on the data environment, the data acquisition, and the hand gesture presentation— three crucial components of the vision-based HGR system. We also looked at how well the vision-based HGR system performed regarding recognition accuracy. [ABSTRACT FROM AUTHOR]

Various machine learning algorithms are developed with an aim to create precise and trustworthy models and extract knowledge from data sources. Deep expertise in the field of machine learning is required for the challenging task of choosing the right algorithms for a specific dataset. There is no single algorithm that outperforms all others across all applications and different datasets. The difficulty of choosing an appropriate algorithm for a specific task in specific domain is related to the properties of the dataset. Properties of the dataset are measured through meta-features. Meta-features describe task and can provide explanation how one machine learning approach outperforms other algorithms on a given dataset. Learning about the effectiveness of learning algorithms, or meta-learning was developed to deal with this issue. Focus is required because previous research papers have not successfully identified meta-features in particular domains. In this research, we have evaluated various meta-feature characterization methodologies and have concentrated on basic meta-features. Business domain data is in the focus of this paper. We computed basic (general) meta-features and illustrated several use cases for their applications. [ABSTRACT FROM AUTHOR]

The paper proposes a novel approach for recognizing and classifying gym exercises using human pose detection with Mediapipe and LSTM, followed by matching the accuracy of the exercises using the Hungarian algorithm and cosine similarity. The proposed approach involves the use of a combination of cutting-edge technologies, including computer vision and deep learning, to recognize and classify a range of gym exercises accurately and efficiently. This study presents several advantages of this approach, including the ability to leverage temporal information, robustness to variations in lighting, camera angle, and occlusions, and the ability to handle a wide variety of gym exercises. Experimental results demonstrate that the proposed approach achieves high accuracy and outperforms existing methods for gym exercise recognition and classification. The paper also compares the proposed approach with existing techniques, highlighting its strengths and cost-effectiveness. It concludes by outlining the methodology, including dataset collection and pre-processing, landmark extraction, LSTM model training, exercise classification, and similarity analysis. Finally, the paper discusses potential future research areas in this field. [ABSTRACT FROM AUTHOR]

Natural disasters cause devastation, chaos, destruction, death, displacement, and much more when they occur. With the advancement of recent technologies especially in artificial intelligence and deep learning, the severity and impact of natural disasters could be reduced by predicting their occurrence. In this research, the flood detection (using images) system is developed using the proposed architecture of the Deep Convolutional Neural Network (CNN). In this paper, the main aim is to propose an automated System for precision disaster detection, in particular flood detection. The intelligent system will monitor vast problematic areas that are dubbed flood risk areas during the season. The intelligent automated system will systematically take images with real-time processing to detect floods. If a flood is detected, the system will automatically communicate to the ground station with the early warning to issue an early warning to all the inhabitants in the path of the flood. Since this paper will prove a concept, the other main contribution of this paper is a new dataset consisting of 9000 images collected from various sources on the Internet, labeled, and preprocessed. The Deep CNN architectures were trained and tested by constructing a new dataset of more than 9000 labeled images with flood and non-flood collected from different sources. Several experiments are performed with changing CNN architectural design parameters in order to get the best recognition rates. Experimental results show that 92.50% accuracy was achieved using GoogleNet. The second highest performance was achieved using AlexNet with an accuracy of 92.20%. [ABSTRACT FROM AUTHOR]

Road traffic by vehicles is continuously increasing through the world and they lead to terrible traffic jamming and significant delay in emergency services at cross roads. The vehicles stuck in that traffic jam may reach their destination with more delay which was lead to loss of lives or a property in some cases for example, emergency vehicles when stuck in the traffic jam like ambulance, fire engine vehicles, police jeeps, etc. Therefore it becomes most significant to control the vehicle traffic on the roads so as to acquire the effective control over the vehicular traffic based on the priority that may leads to saving the life of people in emergency situations by providing delay less services. This paper presents a survey on various methods and systems developed for the traffic classification according to the priority of emergency vehicle and for scheduling the vehicles accordingly in traffic. In this paper, different deep learning methods explored by various works are presented for effective increasing of vehicles creating traffic on the route of emergency vehicle. Finally, the works presented in this paper are compared and result analysis can be illustrated. [ABSTRACT FROM AUTHOR]

The paper presented the various techniques that can be applied for vehicle number plate identification. The extensive literature survey has been carried out on already presented technologies. Out of these technologies, three technologies have been selected and explained in this paper, the technologies are: OCR Method, Deep Learning Method and OpenCV Method. The comparison of three different technologies also available for vehicle number plates identification has been carried out on the basis of accuracy, complexity and cost. [ABSTRACT FROM AUTHOR]

In the diagnosis and treatment of a patient, blood cell detection and classification is a critical activity. Skilled lab technicians manually inspect the blood cells in a conventional method, which is time-consuming and error-prone. With the recent advancements in image processing, deep learning and object detection techniques has acquired larger attention for automating the procedure of blood cell detection and classification. This article focuses on how image processing and machine learning can be used to morphologically characterise and recognise cell pictures collected from peripheral blood smears. Image processing techniques for blood cell detection are typically based on thresholding, segmentation and morphological operations. Machine learning algorithms can learn from data and adapt to new conditions allowing for more accurate and robust blood cell detection. Deep learning algorithms can learn to extract relevant features from raw image data, eliminating the need for manual feature engineering. Deep learning methodology is superior to traditional image processing methods in literature. This paper also focuses on typical generic architectures for object detection covering one stage as well as two stage detectors to improve the detection performance further. Multiple approaches to microscopic blood cells examinations are analysed and compared using various performance metrics. [ABSTRACT FROM AUTHOR]

Image clustering is an important technique in computer vision and image analysis that involves grouping similar imagestogether based on their visual features. It can help to improve the accuracy of visual search and retrieval systems, reduce storage requirements for large image datasets, and identify patterns and trends in image data. This paper discusses an approach to cluster images using K Means clustering algorithm. The purpose of this study is to investigate the effectiveness of k-means clusteringin grouping similar images based on their visual features. The feature extraction is using the ResNet50 architecture, a popular deep learning architecture that has achieved state-of-the-art performance on a variety of computer vision tasks, including image classification, object detection, and semantic segmentation. ResNet consists of multiple residual blocks, each containing several convolutional layers and shortcut connections that enable the network to learn a residual mapping from the input to the output. The architecture is modified in such a way that the output from the flatten layer is taken as the feature of each image. The flatten layer of ResNet50 result 2048 nodes as output. These 2048 values are the features of each image. The feature list is created from ResNet50 model and is saved as a feature file; this detail is used for clustering the values into different clusters. The experiment was conducted on a dataset of 12000 images, and the results show that the k-means algorithm is effective in clustering similar images together. The study also compares the performance of different k values on the dataset using elbow method. [ABSTRACT FROM AUTHOR]

Ayurveda is oldest method to make essential medicines. The medicinal plants have been played a crucial role for curing the various diseases throughout the human history. Although the identification of medicinal plant is very complex task. The factors such as surrounding, growth phases, soil types and lighting difference are the causes for making the identification process more complex. The medicinal plants are classified according to their colour, shape, texture, etc. In this paper we explore various deep learning algorithms to classify the medicinal plants according to their morphological features. The Convolution Neural Network (CNN) based techniques are one of the efficient techniques for identification of medicinal plant using image dataset. The dataset is consist of large number of images of leaf of medicinal plants. Hence the leaves of the medicinal plant are classified based on their characteristics. Computer vision also plays an important role for identification of types of eaves as we as for collecting dataset. Thus, the deep learning technologies helps to provide an accurate and efficient result for the classifying the medicinal plants. [ABSTRACT FROM AUTHOR]

Fault detection is a critical stage in power systems, requiring swift and accurate identification. This paper introduces an Artificial Neural Network (ANN) tailored for fault identification, utilizing a deep learning approach. The proposed model marks a pioneering effort, employing ANN for fault identification using raw and sampled data from three-phase current and voltage signals collected through simulation from a 14-bus system across various fault and no-fault classes. Performance evaluation metrics such as accuracy, sensitivity, specificity, precision, and F1 score are employed to assess the model's efficacy. Results demonstrate that the ANN model achieves exceptional accuracy and efficiency in fault detection, showcasing its potential as a pivotal tool for enhancing the reliability and performance of electrical power systems. Notably, the suggested model attains a remarkable 99% accuracy across all tested fault scenarios, F1Score 0.99 for fault and non-fault detection with closely matched testing and training loss curves. [ABSTRACT FROM AUTHOR]

In the landscape of predictive analytics, commendable success has been garnered by Machine Learning models in the accurate estimation of age. The critical ability to discern deviations from typical age progression, which may signify accelerated aging or variations from the norm, underscores the demand for diagnostic techniques that are both efficient and accurate. Over time, contributions have been presented for the task that makes use of data-driven techniques for modeling the system, with recent attention being centered on the popularity of deep neural networks, commonly known as deep learning. This review aims to serve as an in-depth analysis of the existing literature on deep learning in age estimation, including a thorough overview of the main deep learning architectures implemented for this end. A quantitative assessment of the collective body of publication research in the field to date is considered. Moreover, the various approaches to age assessment are criticized in the evaluation, indicating the few but significant advantages and disadvantages of each used framework. Unexamined routes as well as really innovative ways of dealing with age estimation in the rapidly changing field are researched in this article. Finally, as a result of such thorough examination, the paper aims to become a very useful and solid resource for the audience of both novice and expert students equally. [ABSTRACT FROM AUTHOR]

The relevance of timely detection of aggressive and suspicious behavior from video surveillance is gaining in importance as surveillance technologies are increasingly used for law enforcement and civilian protection. The biggest difficulty is detecting the presence of something out of the ordinary, which is further complicated by potential high pedestrian loads in a crowded place. In this paper, we will consider various methods for recognizing violent behavior and introduce a detection system by integrating them and separating them into three different subfields. Classification techniques include Stochastic Gradient Descent, Adaptive Boosting, Naive Bayes, Logistic Regression, and Convolutional Neural Network methodologies using both ML and deep learning algorithms. Feature extraction and violence detection techniques are produced by using PCA to drastically minimize the execution time of the technique. The efficiency of the proposed models is evaluated using the RWF-2000 dataset to demonstrate false positive resistance. The Accuracy metric is applied to the false positive to estimate the models' ability to withstand false positives. Thus, this performance benchmark is established with an ACC of 99% on the RWF-2000 dataset to prove the capability of the proposed models. The experimental results of the current research work demonstrate that the proposed CNN model is more generalizable. [ABSTRACT FROM AUTHOR]

Lung diseases are one of the largest problems in public health, and they need fast and accurate diagnostic equipment. Thus, this research offers a unique approach that enhances the classification of lung diseases through deep learning techniques using an ensemble model developed by combining convolutional neural network (CNN) and long short-term memory (LSTM) models. This paper uses a Kaggle-based dataset for diverse lung pathologies like normal, COVID-19-infected scans, tuberculosis, and bacterial lungs. To accomplish an overall performance, the hybrid of CNN and LSTM has been done which resulted in an accuracy of 96.2%. This integrated framework includes both geographical aspects and temporal characteristics of lung disorders that lead to accurate and optimal performance results for lung disorder detection. The proposed model can detect many diseases including COVID-19 and tuberculosis rapidly, thus capable of providing timely intervention. This model has clinical importance due to its potential to transform both diagnosis and standardization concerning lung disease as it is able to detect timely patterns in the images. In addition to proving the applicability of the ensemble CNN-LSTM model, this study opens up opportunities for continued improvement in the early detection and management of lung diseases. In addition, the proposed work has been compared with state-of-the-art lung disease detection models which showed that the ensemble CNN-LSTM model is better than other methods in producing accurate and optimal generalization of the results. [ABSTRACT FROM AUTHOR]

Around the globe, monkeypox has become a disease that spreads quickly, and an outbreak of Monkeypox reported in more than 75 countries. The signs and symptoms of monkeypox disease, which occurred in humans, are very similar to chickenpox and the measles, with very subtle variations like different types of skin blisters etc. Traditional approaches primarily rely, on the texture, shape, feature and colour of the rashes/blisters. The majority of these features which are specified by the problem and are also demonstrated to be complementary in the medical field. This results in a system that is unable to represent high-level problem domain concepts. Till now, various deep learning methods have been implemented to recognize various skin diseases, chest images, COVID-19, tumour cells etc. In this paper, a deep-learning-based approach has been implemented for the recognition of Monkeypox from the different images of various skin diseases. [ABSTRACT FROM AUTHOR]

The timely detection of COVID-19 plays a critical role in its management and control. This research paper presents a novel and effective COVID-19 detection technique that is based on modified DeTrac model. The key notion of this work is to enhance the accuracy of detection rate while lowering its complexity. The novelty of this work lies in Feature Extraction and Feature Selection phases. In the feature extraction phase, two approaches are employed: utilizing a pretrainedAlexNet DL architecture and implementing statistical, textual, and Principal Component Analysis (PCA) techniques. These techniques abstract relevant attributes from CXR images, capturing discriminative information. To enhance the performance and efficiency of the model, an innovative feature selection technique using the Improved Salp Swarm Algorithm (ISSA) is applied. This technique selects crucial features from the extracted feature set, reducing model complexity and improving discriminatory power. The classification phase utilizes an improved DeTracmodel wherein different layers are added in Deep Learning (DL) network architecture, which effectively learns complex and hierarchical representations for accurate classification into three classes of normal, COVID-19, and pneumonic. Extensive experimentation and analysis are conducted to authenticate working of proposed model. The results demonstrate that proposed model outshines standard DeTrac model by achieving an accuracy of 96% for Covid and 95% for Normal and Pneumonic diseases. [ABSTRACT FROM AUTHOR]

Artificial Intelligence (AI) specifies simulation of human intelligence through machines to acquire the ability to act rationally like humans for solving problems. The design of an artificially intelligent system is based on various factors such as natural language processing (NLP), representation of knowledge, automated reasoning, machine and deep learning, computer vision, robotics etc. The area of computer vision (CV) focuses on reproducing parts of the complex human visual system and enables the computer systems to identify, analyze, understand, and process the objects present in an image like the human brain. There are various sub-domains of computer vision, such as object/event detection, object/activity identification, pose/motion estimation, and scene/image reconstruction etc. Object detection is a computer vision task employed to detect the instances of visual objects of a particular class in an image. There are different platforms available for achieving real-time object detection in different applications of intelligent systems. This paper surveys the computer vision implementation tasked across object detection using Machine Learning (ML) and Deep Learning (DL) algorithms for design of intelligent systems. The computer vision toolbox available in MATLAB is also very popular among researchers for its high utilization of advanced features for object detection and its tracking. This study provides a comprehensive summary of state-of-the-art technologies for implementation of computer vision and discusses the perspectives of applications in intelligent systems achieving real-time object detection in different domains. [ABSTRACT FROM AUTHOR]

Hyperspectral Imagery (HSI) often suffers from degradation caused by different types of noise, including Gaussian noise, impulse noise, and stripe noise. Restoring the quality of HSI is a complex task, particularly due to the high dimensionality of HSI datasets. In recent years, there has been a growing utilization of low-rank tensor-based and deep learning-based approaches for denoising HSI data, especially in scenarios involving mixed noise. In this research paper, we assess the performance of two low-rank-based methods and one deep learning-based method when applied to benchmark HSI datasets with a focus on the removal of mixed noise. We quantitatively evaluate the restored datasets using a range of quality metrics and subject them to qualitative visual assessment. Our experimental findings reveal that the low-rank-based methods exhibit promising results in effectively mitigating mixed noise in HSI data. These methods offer a viable solution for addressing the challenge of noise removal in hyperspectral imagery, contributing to improved data quality and usability. [ABSTRACT FROM AUTHOR]

Gender classification refers to the process of categorizing individuals into one of two gender categories: male or female, typically based on observable characteristics or information. This classification can be done through various methods, including biological and social. In recent years, gender classification has become a topic of increasing interest and debate due to evolving societal understandings of gender. The current survey will study the connection between language use and gender to categorize gender automatically based on text and linguistic style. It provides an in-depth analysis of gender classification based on linguistic patterns in written text. It explores the relationship between linguistic patterns and gender classification, highlighting the various approaches, challenges, and future directions in this field. It also covers various datasets that classify people by gender, including official papers, emails, and social media messages. This survey divides the selected studies into three parts: handwritten, names, and text. However, the most focused part is text based on linguistic analysis. The findings show that the most used dataset is Twitter. Many studies use English, Arabic, and other languages such as Portuguese, Chinese, Spanish, Russian, Brazilian, and German. Moreover, the feature frequently used in studies is the Bag of Words (BOW). Also, the methodology used in many studies is machine learning techniques; however, few use deep learning. Finally, the important metrics are accuracy and F1-score. [ABSTRACT FROM AUTHOR]

In the rapidly evolving landscape of digital communication, ensuring the reliability and security of networks is paramount, with the escalating volume and complexity of network traffic necessitating robust anomaly detection mechanisms to safeguard against potential threats and disruptions. This paper introduces a novel hybrid deep learning system designed for effective network anomaly detection, employing a two-stage approach. The first stage utilizes a Convolutional Neural Network (CNN) as a binary classifier to identify broad anomalous patterns in network traffic. Subsequently, the second stage employs a Deep Neural Network (DNN) as a multi-class classifier to further categorize the detected anomalies into specific classes. Notably, a unique feedback mechanism is incorporated into the system, leveraging false results, including false negatives and false positives, to iteratively refine the model through retraining and reclassification, thereby mitigating false results. An innovative aspect of the proposed approach lies in the incorporation of a feedback loop that addresses false results. False negatives and false positives identified during the classification process are used to augment the training dataset, allowing the system to learn from its own mistakes. The model is then retrained, and the classification process is repeated, resulting in a more resilient and adaptive anomaly detection system. To evaluate the performance of the proposed hybrid system, extensive experiments are conducted using the CICIDS2017 dataset, a widely recognized benchmark in the field of cybersecurity. Comparative analyses with existing methods showcase the superior capabilities of the hybrid system in terms of accuracy, false positive reduction, and adaptability to evolving network threats. The results indicate that the hybrid deep learning system not only outperforms traditional approaches but also exhibits an enhanced ability to learn and adapt from misclassifications, thereby continually improving its performance over time. In conclusion, the presented two-stage hybrid deep learning system, coupled with the adaptive feedback mechanism, represents a significant advancement in network anomaly detection, promising heightened accuracy and robustness in the ever-evolving landscape of cybersecurity. [ABSTRACT FROM AUTHOR]

With the rapid development of society and the intensification of employment competitiveness, the mental health problems of college students have attracted widespread attention. The rapid increase in mental health problems among college students has brought great challenges to college students and university education management institutions. Therefore, the development of an effective college student mental health early warning algorithm and corresponding systems to support mental health management has become an urgent problem to be solved. In view of the above problems, this paper proposes a deep learning algorithm for college students' mental health early warning model, which collects a large amount of mental health data, combines the advantages of distributed multi-agent systems, divides the data into multiple data subsets, and each agent learns and extracts the characteristics of potential mental health problems according to the corresponding data subset, and finally, the final early warning scheme is obtained by integrating the learning results of each agent. Experimental results show that the algorithm has high accuracy and efficiency, can accurately identify potential mental health problems of college students, and warn in advance. Compared with traditional prediction methods, the deep learning algorithm for college students' mental health warning has better scalability when processing large-scale data. In addition, deep learning algorithms can quickly adapt to the differences between different time periods and individuals, providing more comprehensive support for college students' mental health management. [ABSTRACT FROM AUTHOR]

With the rapid development of artificial intelligence, basic tasks such as automatic sample classification and object detection of computers have also developed rapidly. How to obtain the required information quickly, accurately and comprehensively from massive information has become the research hotspots. However, traditional classification detection methods have some limitations, such as high dependence on feature extraction and model selection, which greatly reduces the detection efficiency and accuracy. Therefore, this paper proposes a deep learning algorithm for combinatorial optimization classification detection and analysis. Firstly, statistical theory is used to process the computer target sample information, and the indicators are divided automatic classification to reduce the factors in classification detection. Then, statistical theory automatically classifies computer target samples, forms an automated classification scheme, and comprehensively analyzes the classification detection results. The results show that the accuracy and classification detection time of the computer target sample automated classification of the deep learning algorithm are better than the traditional classification detection method under the condition that the processing standard is fixed. [ABSTRACT FROM AUTHOR]

Housing price prediction and increasing research-based studies using Machine Learning and Deep Learning are popular. Some researchers have not considered all-inclusive information that affects the price prediction. Motive of this paper is to show variance score, measure of errors and predicting the feature price of the house performed by two different models. Here we are using Machine learning (ML) and Deep Learning models to calculate the metric. By using the Keras-ANN model performs both variance score and predicting new price of the house. It inputs a single data of a house in Boston city and predicts the new price based on the features of house. Model Evaluation Measures is used to compare the performance of both models. The result describes the measure of variability in evaluation scores. [ABSTRACT FROM AUTHOR]

The aim of this research paper is to compare the accuracy of a novel Visual Geometry Group-19 deep learning model and a CNN in order to detect bone fractures in the upper extremity regions of the hands. The two groups involved in this investigation are the Novel Visual Geometry Group-19 deep learning model and the CNN. This research makes use of an open source dataset titled "Bone Fracture Detection Using X-Rays" obtained from Kaggle. The sample size of the dataset consists of 9463 x-ray images. Test and train sets were constructed from the dataset. The training dataset comprises 8987 images, while the test dataset comprises 633 images. Deliberate consideration was given to the iteration sample size with a 95% confidence interval and a G power of 80%. Ten iterations of the experiment were conducted utilizing the aforementioned models. The analysis of the bone fracture dataset utilizing x-rays revealed that the convolutional neural network (CNN)achieved a 72.50% accuracy rate, while the innovative Visual Geometry Group-19 deep learning model achieved a 97.01% accuracy rate. The t-test indicates that the novel Visual Geometry Group-19 deep learning model exhibited greater statistical significance than the CNN (CNN), as evidenced by its p<0.001 (2-tailed) value. The results indicate that the innovative Visual Geometry Group-19 method is effective at detecting bone fractures in the upper extremities and is practical for early fracture prediction. [ABSTRACT FROM AUTHOR]

The abstract examines the need for prompt and accurate diagnosis of Autism Spectrum Disorder (ASD), a neurodevelopmental disorder characterized by difficulties in social communication and repetitive behaviors. Early diagnosis is crucial in facilitating appropriate intervention and support, thereby substantially improving the potential for children's social and cognitive development. The conventional diagnostic approaches for ASD are often characterized by lengthy durations, high expenses, and susceptibility to inconsistencies across evaluators. These challenges are primarily due to the complex nature of the illness, which requires a complete evaluation encompassing several aspects such as behavior, neurology, and other relevant characteristics. Given the obstacles mentioned above, applying deep learning methodologies has emerged as a potentially fruitful approach for diagnosing ASD. These techniques demonstrate exceptional proficiency in identifying significant patterns and characteristics from diverse datasets, making a valuable contribution to developing more accurate and effective diagnostic models. This comprehensive literature review examines recent studies on the diagnosis of ASD, particularly emphasizing the various types of datasets employed in deep learning-based methods. The study covers various datasets that include behavioral data, capturing intricate details of social interaction and communication patterns, neuroimaging data that provide insights into the structure and function of the brain, and datasets that incorporate genetic and clinical assessments. The paper provides a more in-depth analysis of the challenges faced while acquiring datasets, which include issues related to data reliability, the adequacy of sample sizes, and the variability present within the datasets. [ABSTRACT FROM AUTHOR]

Since many decades ago, there has been a large increase in the use of neonatal hip ultrasound imaging, making it simpler to identify hip developmental anomalies (DDH). In this paper, a novel strategy for data enhancement is proposed. It entails the use of an improved Bayesian non-local median filtering to reduce speckle noise (OBNLM). After already being supplemented with information from the OBNLM, the intended CNN system's accuracy is improved from 92.29 percent to 97.70 percent. This novel method for automatically classifying DDH detects dysplastic newborn hips with a high accuracy rate, which may assist evaluators boost the success rate of their assessments. For a fully automated evaluation of the scan effectiveness of three-dimensional ultrasound volumes, we provide a deep learning approach. We trained a Convolutional Neural Network to determine whether or not 3D Ultrasound (3DUS) hip pictures are diagnostically adequate (CNN). 2187 3DUS images were included in Dataset 1 (DS1), and the quality of the scan was evaluated by a single reader using a scale from 1 (the lowest standard) to 5 (the highest quality) (optimal quality). Dataset 2 consists of 107 3DUS photos that were evaluated semi-quantitatively by four users using a 10-point scale (DS2). It was determined whether or not these datasets were suitable for testing the efficacy of this AI method. The majority of the Intraclass Correlation Coefficient (ICC) and Cohen's kappa coefficient showed that the AI method agreed well with expert readings, and the AI method's extremely accurate estimates on both datasets as a binary classifier (adequate/inadequate) (DS1 accuracy = 96% and DS2 accuracy = 91%) were consistent with those obtained by the experts (K). Our AI-powered technology can be used as a screening tool during ultrasound imaging or post-processing to ensure a higher quality scan and a more precise assessment of the baby's hip. Additionally, the characteristics that the convolutional layers gain show that FNet can concentrate on crucial characteristics, which is useful for accurately reconstructing the femoral head anatomy. In conclusion, the recommended approach is effective in appropriately segmenting the femoral head and providing recommendations for assessing hip development abnormalities. [ABSTRACT FROM AUTHOR]

The quantity and quality of food produced are often threatened by plant diseases, which can have a devastating impact. Therefore, early diagnosis of plant diseases is often crucial to prevent production losses. The development of deep learning modelshas opened up new avenues for achieving high accuracy in detecting and mitigating crop diseases. This paper proposes a CNN models for accurately classifying leaf diseases in crops, specifically focusing on rice, corn, and wheat plants. The proposed method utilizes DenseNet169 and InceptionV3 as the base models respectively trained on corn, rice, and wheat disease datasets. The proposed method has successfully achieved an accuracy of 99.89% using DenseNet169 and 99.25% using InceptionV3 for corn leaf disease dataset. The proposed method also obtained an accuracy of 98.52% and 98.58% for the wheat crop disease using DenseNet169 and InceptionV3 models respectively. For the Rice crop, rice dataset is separated into major and minor diseasesin order to comprehend the variety of diseases and obtained an accuracy of 99.18% and 91.77% for major and minor rice crops using DenseNet169 variants respectively. Overall, the proposed approach proves to be successful in accurately classifying diseasesacross rice, corn, and wheat crop diseases, providing reliable and detailed predictions. [ABSTRACT FROM AUTHOR]

Object detection is one of the most challenging problems in computer vision. Detecting an object from an image or video finds applications in different fields like security, medical, automated vehicle systems, etc. Locating instances from a real-world image requires improved accuracy and low latency. Deep learning (DL) techniques have emerged as a powerful tool for training the system and detecting instances. However, DL-based object detection for real-time applications requires a lot of computational resources. Field Programming Gate Array (FPGA) is found to be more promising to give optimized results for object detection using deep learning. This paper presents a detailed study of various DL-based models using different datasets like MS-COCO, PASCAL VOC, etc for object detection. It also provides good insight for selecting an algorithm for an object detection application with FPGA architecture. The study shows that the advancement in software and hardware technologies enables accurate and efficient detection systems for a wide range of applications. [ABSTRACT FROM AUTHOR]