Body posture dynamics have garnered significant attention in recent years due to their critical role in understanding the emotional states conveyed through human movements during social interactions. Emotions are typically expressed through facial expressions, voice, gait, posture, and overall body dynamics. Among these, body posture provides subtle yet essential cues about emotional states. However, predicting an individual’s gait and posture dynamics poses challenges, given the complexity of human body movement, which involves numerous degrees of freedom compared to facial expressions. Moreover, unlike static facial expressions, body dynamics are inherently fluid and continuously evolving. This paper presents an effective method for recognizing 17 micro-emotions by analyzing kinematic features from the GEMEP dataset using video-based motion capture. We specifically focus on upper body posture dynamics (skeleton points and angle), capturing movement patterns and their dynamic range over time. Our approach addresses the complexity of recognizing emotions from posture and gait by focusing on key elements of kinematic gesture analysis. The experimental results demonstrate the effectiveness of the proposed model, achieving a high accuracy rate of 91.48% for angle metric + DNN and 93.89% for distance + DNN on the GEMEP dataset using a deep neural network (DNN). These findings highlight the potential for our model to advance posture-based emotion recognition, particularly in applications where human body dynamics distance and angle are key indicators of emotional states.

Relating specifically to human–computer interaction (HCI), computer vision research has placed a substantial emphasis on intelligent emotion recognition in recent years. The primary emphasis lies in investigating speech aspects and bodily motions, while the knowledge of recognizing emotions from facial expressions remains relatively unexplored. Automated facial emotion detection allows a machine to assess and understand a person's emotional state, allowing the system to predict intent by analyzing facial expressions. Therefore, this research provides a novel parameter selection strategy using swarm intelligence and a fitness function for intelligent recognition of micro emotions. This paper presents a novel method based on geometric visual representation obtained from facial landmark points. We employ the Deep Neural Networks (DNN) model to analyze the input features from the normalized angle and distance values derived from these landmarks. The results of the experiments show that Particle Swarm Optimization (PSO) worked very well by using only a few carefully chosen features. The method achieved a recognition success rate of 98.76% on the MUG dataset and 97.79% on the GEMEP datasets.

One interesting aspect of distributed systems is cloud computing. It offers its services on an as-needed, pay-per-use basis. Research challenges in cloud computing include the scheduling and balance of tasks. Scheduling tasks entails assigning them to available resources (Virtual Machines), whereas load balancing is spreading out the work over multiple available resources. We offer a job scheduling technique that takes into account anticipated demands on compute nodes. We begin by investigating the reasons for the disparity in workloads and the viability of redistributing computing resources. All of the complex simulations are run in clouds. The process consisted of two distinct parts. The simulation begins with a randomly generated workload. Second, we combine the workload prediction model with an application and workload-aware scheduling algorithm (AAWAS) [14]. We present a parallel job scheduling approach using computational node workload prediction AAWAS to simplify the AAWAS algorithm. Experimental results reveal that AAWA is superior to other algorithms in terms of minimizing makespan and optimizing resource utilization, and the proposed method is compared to FCFS, SJF, Min-Min, and EDF. To make Task scheduling in compute clouds faces various challenges due to the highly dynamic environment, such as availability, access policies, security, and reliability of computing resources. Furthermore, the collaborative nature of cloud computing projects on the internet adds complexity to task scheduling. From simulation results, it is evident that our proposed scheduler AAWAS outperforms A2C, DQN, and PPO across the board by 4.1%, 2.6%, and 3.4%, respectively. Because AAWS relies on policy for learning rather than Q-value, its superiority can be defended. Several environmental factors affect AAWAS and contribute to its consistent learning process and the high chance of accurate predictions. As a result, AAWASC may learn from experiences independently of target networks, and this helps to better generalize the network.

Alzheimer's Disease is a progressive neurodegenerative condition distinguished by a steady deterioration in cognitive abilities. Given the empirical inquiries conducted, it is imperative to take age into account as a fundamental criterion when selecting participants. Younger individuals are more vulnerable to the transient aspect compared to those who are older. The current work focused on selecting young-onset subjects, as early diagnosis can significantly impact patients' lives. The identification of this disease at an early stage using traditional means poses significant difficulties. Deep Learning (DL) has emerged as a highly effective approach for enhancing the performance of diagnostic processes and boosting forecast accuracy. The research utilized deep learning techniques and neuroimaging approaches to autonomously identify and categorize the condition based on its phases of mild cognitive impairment. The suggested study is conducted through a series of three sequential phases. Prior to further analysis, the 3D input visuals need to undergo image preprocessing with Weiner filtering in order to remove noise and smooth the image. Following, Transfer learning models are utilized to extract features, which are then compressed through the use of cascaded Auto Encoders. The ultimate stage involves the utilization of a fine tuned Deep Neural Network (DNN) for categorizing of the phases of AD into five classes. The combination of the ResNet-18 and sparse autoencoder with the deep neural network model yielded a remarkable accuracy rate of 95.62

The wireless networks do not consider ultra-high reliability and low latency (URLL) as the primary factors. There is a delay-tolerant, non-critical content and services that are based on human-centric communication in mobile networks. The focus of the wireless networks is to provide boosted rate in data speed and increasing network coverage by using the best networking approach. The wireless communication device is primarily concerned with determining the entity’s status, by offering enhanced wireless connection assistance, and thereby raising the commodity’s status. Additionally, there will be occurrence of traffic in an infrequent manner. This adds pressure for the delivery of packets that are large or medium-sized at low latency. This is not provided by the current wireless technology system’s state of the art. This is a vital factor that challenges the future development of the wireless system for communication. Thus, to have high reliability with the irregular traffic patterns, there is a need for the development of novel and innovative technologies from physical to the network layer in the wireless communication systems. In this paper, technologies like multiple-access scheme, synchronization, multiplexing, full-duplex transmission, and channel code design are introduced for spectrally systematic ultra-high reliability and low latency communication (URLLC) in the physical layer. The techniques developed for the purpose of monitoring and controlling plant growth are called precision agricultural techniques. These techniques use actuators and sensor networks to monitor and control plant growth. Precision agriculture involves smart irrigation and fertilizer monitoring. This paper also develops a new precision agriculture technique using wireless sensor network (WSN). The WSN uses the Message Queuing Telemetry Transport (MQTT) protocol. This WSN uses IoT technology for connecting devices with the sensors for the purpose of monitoring, collecting, and distributing the data. The IoT integrates with cloud computing to subdue the limitations of IoT in monitoring and controlling systems. Thus, this paper involves the development of the UHLL communication system that can be used in precision agriculture using WSNs.

The rapid advancement of innovation has sparked significant improvements in the field of music creation, with chord progression being one of the notable areas affected. Many musicians face challenges in playing instruments like the piano by relying solely on auditory perception. Therefore, the need for a chord generator arises to assist them in this endeavor. Mastering music composition demands a substantial investment of time, particularly for novice musicians. To alleviate the time-consuming nature of this task, we explore the practicality of automated chord progression generation by analyzing audio files and converting them into strings. First, the music or audio file will be converted into strings using the frequency of that record. Secondly, these strings will be employed to identify the individual musical notes corresponding to those specific strings. Ultimately, the best chord for the music will be generated, making the process more efficient for musicians.

Alzheimer's Disease (AD) is a progressive neurological disease. Early diagnosis of this illness using conventional methods is very challenging. Deep Learning (DL) is one of the finest solutions for improving diagnostic procedures' performance and forecast accuracy. The disease's widespread distribution and elevated mortality rate demonstrate its significance in the older-onset and younger-onset age groups. In light of research investigations, it is vital to consider age as one of the key criteria when choosing the subjects. The younger subjects are more susceptible to the perishable side than the older onset. The proposed investigation concentrated on the younger onset. The research used deep learning models and neuroimages to diagnose and categorize the disease at its early stages automatically. The proposed work is executed in three steps. The 3D input images must first undergo image preprocessing using Weiner filtering and Contrast Limited Adaptive Histogram Equalization (CLAHE) methods. The Transfer Learning (TL) models extract features, which are subsequently compressed using cascaded Auto Encoders (AE). The final phase entails using a Deep Neural Network (DNN) to classify the phases of AD. The model was trained and tested to classify the five stages of AD. The ensemble ResNet-18 and sparse autoencoder with DNN model achieved an accuracy of 98.54%. The method is compared to state-of-the-art approaches to validate its efficacy and performance.

A video recommendation framework for e-commerce clients is proposed using the collaborative filtering (CF) process. One of the most important features of the CF algorithm is its scalability. To avoid the issue, a hybrid model-based collaborative filtering approach is proposed. KL Divergence was developed to address the CF technique’s scalability problem. The clustering with enhanced sqrt-cosine similarity Recommender scheme is proposed. For successful clustering, Kullback–Leibler Divergence-based Fuzzy C-Means clustering is suggested, with the aim of focusing on greater accuracy during movie recommendation.The proposed scheme is viewed as a trustworthy contribution that significantly improves the ability of movie recommendation by virtue of the KL divergence-based Fuzzy C-Means clustering mechanism and enhanced sqrt-cosine similarity. The proposed scheme highlighted and addressed the critical role of the KL divergence-based cluster ensemble factor in improving clustering stability and robustness. For prediction, the enhanced sqrt-cosine similarity was used to calculate successful related neighbor users. The performance of Recommendation is improved when KLD-FCM is combined with improved sqrt-cosine similarity.The proposed scheme’s empirical work on the Movielens dataset in terms of MAE, RMSE, SD, and Recall were found to be superior in recommendation accuracy compared to traditional approaches and some non-clustering based methods recommended for study. With the specified number of clusters, it is capable of providing accurate and customized movie recommendation systems.

Landslides are a natural hazard that is unpredictable, but we can prevent them. The Landslide Susceptibility Index reduces the uncertainty of living with landslides significantly. Planning and managing landslide-prone areas is critical. Using the most optimistic deep neural network techniques, the proposed work classifies and analyses the severity of the landslide. The selected experimental study area is Kerala’s Idukki district. A total of 3363 points were considered for this experiment using historic landslide points, field surveys, and literature searches. The primary triggering factors slope degree, slope aspect, elevation (altitude), normalized difference vegetation index (NDVI), and distance from road, lithology, and rainfall are considered. A landslide susceptibility map was generated using the Arc geographic information system (GIS) tool for all the triggering factors using frequency ratio method, Shannon entropy method, Relative effect method, and fuzzy logic method. A new Deep Neural Network (DNN) framework has been developed for the multiclass classification and prediction of landslide hazard zones as low, moderate, high, and very high. Existing works are only uses statistical methods, but the proposed work has used DNN to predict landslide severity at four different level even for semi data with over accurately. The training data for deep learning model are generated using the Sentinel Satellite images and field survey. The label for training the data are generated from the Landslide Susceptibility Index which are generated from statistical method. Among all the statistical method generated data the Shannon Entropy data is the most accurate of the four statistical methods achieves 99.16%, accuracy. The frequency ratio method based data achieves 97.08% accuracy, the relative effect method based data achieves 92.72% accuracy, and the Fuzzy logic method based data achieves 86.60% accuracy.

Aerial image-based target object detection has several glitches such as low accuracy in multi-scale target detection locations, slow detection, missed targets, and misprediction of targets. To solve this problem, this paper proposes an improved You Only Look Once (YOLO) algorithm from the viewpoint of model efficiency using target box dimension clustering, classification of the pre-trained network, multi-scale detection training, and changing the screening rules of the candidate box. This modified approach has the potential to be better adapted to the positioning task. The aerial image of the unmanned aerial vehicle (UAV) can be positioned to the target area in real-time, and the projection relation can convert the latitude and longitude of the UAV. The results proved to be more effective; notably, the average accuracy of the detection network in the aerial image of the target area detection tasks increased to 79.5%. The aerial images containing the target area are considered to experiment with the flight simulation to verify its network positioning accuracy rate and were found to be greater than 84%. This proposed model can be effectively used for real-time target detection for multi-scale targets with reduced misprediction rate due to its superior accuracy.

Alzheimer’s disease (AD) is a type of neuron disease; its nature causes the brain cells to degenerate and die. It’s a progressive disorder. It’s an incurable disease and develops memory impairment as it progresses. The precise diagnosis of AD plays a vital role in the patient’s health care, especially at its initial stage. The early detection of the disorder can help the patient get proper treatment and prevent further irreversible damage to the brain. This paper focuses on a comprehensive study of computer-aided diagnosis of AD by Convolution Neural Network (CNN). Many researchers have performed various ways through CNN for predicting AD at its dawn stage. The comprehensive study of paper shows various algorithms in CNN for early diagnosis of the disease through neuroimaging biomarkers

Recently, increasing attention in the field of gesture recognition, has become a key strategy in analyzing the emotional states of human body movements for social communication. Most real-life scenarios include identifying emotions from facial expressions, vocal synthesis, hand recognition and body gestures. The body posture powerfully conveys the micro emotions of a person in depth. The prediction of human - gait is significantly harder, because the pattern of the human pose estimation has additional degrees of self-determination than the facial emotions, and the overall shape varies robustly during the articulated motion. In this paper, we propose a novel method to recognize 17 different micro emotions from GEMEP dataset based on human upper body gestures dynamics features extracted from the abstract representations of patterns from videos. In the experimental results, KNN exhibit the proposed architecture's effectiveness with an accuracy rate of 97.1% for the GEMEP dataset, 95.2% for SVM, 51.6% for Decision Tree and 49.7% Naive Bayes, respectively.

Alzheimer disease (AD) is an incurable, irreversible brain disorder. It impairs thinking capacity and memory loss. Computer-aided diagnosis techniques with image retrieval have developed a new potential in magnetic resonance imaging, which helps to retrieve relevant images and train to detect AD and its stages. Recently, advanced machine learning techniques have successfully exhibited high scale performances in numerous fields. This paper proposed four machine learning techniques such as Support Vector Machine (SVM), K-Nearest Neighbour (K-NN), Naïve Bayes, and Decision Tree using Brain MRI to identify AD stages. The models encompassed scattering wavelet transform for extracting the relevant features from MRI. While most of the existing techniques focus on binary classification, the current work focused on multi-class classification by classifying the stages of Alzheimer disease, namely healthy controls, very mild AD, mild AD and moderate. The SVM classifiers obtained a superior performance with an average accuracy of 98.10

Facial analysis is an active research topic in examining the emotional state of humans over the past few decades. It is still a challenging task in computer vision due to its high intra-class variation, head pose, suitable environment conditions like lighting and illumination factors in behaviour prediction and recommendation systems. This paper proposes a novel facial emotion representation approach based on dense descriptors for recognizing facial dynamics on image sequences. Initially, the face is detected using the Haar cascade classifer to extract the temporal information from the facial frame by applying a scale invariant feature transform by combining a bag of visual words. Later, the extracted high-level features are fed to machine learning algorithms to classify the seven emotions from the MUG dataset. The proposed dense SIFT clustering performance was evaluated on four different machine learning algorithms and achieved a high rate of recognition accuracy in all classes. In the experimental results, K-NN exhibits the proposed architecture's effectiveness with an accuracy rate of 91.8% for the MUG dataset, 89% for SVM, 87.6% for Naive Bayes, and 85.7% decision tree, respectively.

In Nilgiris mountain regions, landslide is one of the most severe hazards, damages natural resources, and disturbs the ecosystem and environment. It creates a substantial loss to agriculture, forest, wealth and scratches major architectural constructions in the mountainous region. This proposed work aims to distinguish the range of the spatial vulnerability of landslide using ensembled machine learning models, including AdaBoost, random forest algorithms, and gradient boosting decision tree algorithms. Seven conditioning factors namely slope angle, slope aspect, land use, geomorphic features, distance from road, drainage density, and lineament density and a total of 272 landslide historical locations with a ratio of 70/30 were used to construct the spatial database. Estimate the efficacy of the ensembled models using evaluation metrics like precision, recall, F1-score, MAE, MCE, MSE, Kappa and AUC score. Results concluded that the land-use factor, vegetable crop area, is prone to high vulnerability zone in the study area. Additionally, results show that among all the ensembled machine learning models, the AdaBoost algorithm’s classification gives the highest accuracy value of 88.96%. However, the gradient boost decision classifier also outperformed with an accuracy value of 78.43%, and the random forest algorithm gives 72.61%.

In recent years, human action recognition is modeled as a spatial-temporal video volume. Such aspects have recently expanded greatly due to their explosively evolving real-world uses, such as visual surveillance, autonomous driving, and entertainment. Specifically, the spatio-temporal interest points (STIPs) approach has been widely and efficiently used in action representation for recognition. In this work, a novel approach based on the STIPs is proposed for action descriptors i.e., Two Dimensional-Difference Intensity Distance Group Pattern (2D-DIDGP) and Three Dimensional-Difference Intensity Distance Group Pattern (3D-DIDGP) for representing and recognizing the human actions in video sequences. Initially, this approach captures the local motion in a video that is invariant to size and shape changes. This approach extends further to build unique and discriminative feature description methods to enhance the action recognition rate. The transformation methods, such as DCT (Discrete cosine transform), DWT (Discrete wavelet transforms), and hybrid DWT+DCT, are utilized. The proposed approach is validated on the UT-Interaction dataset that has been extensively studied by past researchers. Then, the classification methods, such as Support Vector Machines (SVM) and Random Forest (RF) classifiers, are exploited. From the observed results, it is perceived that the proposed descriptors especially the DIDGP based descriptor yield promising results on action recognition. Notably, the 3D-DIDGP outperforms the state-of-the-art algorithm predominantly.

Human emotion recognition is an active research topic in analysing the emotional state of humans over the past few decades. It is still a challenging task in artificial intelligence and human–computer interaction due to its high intra-class variation. Facial emotion analysis achieved more appreciation in academic and commercial potential challenges mainly in the field of behaviour prediction and recommendation systems. This paper proposes a novel scattering approach for recognizing facial dynamics using image sequences. Initially, we extract the temporal information from the facial frame by applying a saliency map and hyper-complex Fourier transform (HFT). Later the extracted high-level features are fed to the scattering transform method and machine learning algorithms to classify the seven emotions from the MUG dataset. The performance of proposed wavelet scattering network was evaluated on four different machine learning algorithms and achieved a high rate of recognition accuracy in all classes. In the experimental results, K-NN exhibits the proposed architecture's effectiveness with an accuracy rate of 97% for the MUG dataset, 95.7% for SVM, 93.7% for decision tree and 91.2% naive Bayes, respectively.

The size of companies has seen an exponential growth over the years. Corporations recruit anywhere from a few hundred to a few thousand employees every year. With such rates, human resource management in companies is proving to be more and more significant every day. Done manually, HRM is a laborious task, given the sheer quantity of employees. Luckily, over the years, data analytics in HR is emerging as an integral part in corporate operation. Yet, there remain a few tasks that involve human involvement, one of them being selecting candidates that are eligible for a promotion. This paper proposes a solution using decision tree-based machine learning algorithms to learn from past employee records to aid this decision-making process. It explores the usage of two machine learning algorithms, random forest, and XGBoost to predict whether an employee is eligible to receive a promotion or not and determine what factors are responsible for that prediction.

Alzheimer's Disease is progressive dementia that begins with minor memory loss and develops into the complete loss of mental and physical abilities. Memory-related regions of the brain, such as the entorhinal cortex and hippocampus, are the first to be damaged. A person's mental stability is harmed as a result of the severity. Later on, it affects cortical regions that engage with language, logic, and social interaction. Subsequently, it spreads to other parts of the brain, resulting in a substantial reduction in brain volume. Although computer-aided algorithms have achieved significant advances in research, there is still room for improvement in the feasible diagnostic procedure accessible in clinical practice. Deep learning models have gone mainstream in the latest decades due to their superior performance. Compared to typical machine learning approaches, deep models are more accurate in detecting Alzheimer's Disease. To identify the labels as demented or non-demented, the researchers used the Open Access Series of Imaging Studies (OASIS) dataset. The novelty comes in doing extensive research to uncover crucial predictor factors and then selecting a Deep Neural Network (DNN) with five hidden layers and carefully tuned hyper-parameters to achieve exemplary performance. The assertions were supported by evidence of 90.10% correlation accuracy at various iterations and layers.