Defect detection stands as a pivotal quality control measure within the realm of manufacturing. Recent studies have witnessed the successful implementation of defect detection systems across diverse domains, ranging from scrutinizing steel surfaces to assessing fruit grades, and notably, integrating these systems into smart factories. In textile manufacturing, the automated identification of fabric defects has become indispensable, given the intricate nature of fabric textures, where minute imperfections can easily elude human perception. Consequently, relying solely on human inspection proves inadequate for resolving these challenges. Moreover, there has been a notable upswing of interest in applying IoT and machine learning techniques to bolster defect detection processes, owing to their demonstrated reliability. Consequently, the pursuit of defect detection has shifted towards a foundation rooted in Deep Learning. The core objective of this paper revolves around honing fabric defect detection through the utilization of CNN, a renowned deep learning model. This research utilizes CNN based pretrained models for fabric defeat prediction. It takes the popular AlexNet pretrained models for fabric defeat prediction on AITEX Fabric Image Set. The experimental results show that AlexNet provide a good accuracy 93.4% compared with normal CNN model.

Skin disorders are prevalent and difficult medical conditions that impact millions of individuals globally. For these illnesses to be effectively treated and potential complications to be avoided, a prompt and precise diagnosis is essential. Dermatologists and other medical professionals may now diagnose skin diseases and abnormalities more accurately thanks to the development of several automated skin detection technologies. The application of Deep Learning (DL) technology to the diagnosis and categorization of skin disorders has showed considerable potential. These techniques automatically evaluate and categorize skin photos according to patterns, textures, and characteristics using artificial neural networks, more especially convolutional neural networks (CNNs). In this paper, we examine the prediction of automatic skin detection using Mobile-Net model and HAM10000 dataset. This research utilizes the ReduceLROnPlateau function to reduce the learning rate of proposed Mobile-Net model. From this experimental analysis, the proposed Mobile-Net model provides a top 2 accuracy 0.927, top3 accuracy 0.9648 and less validation loss compared with existing mobilenet model.

Ensuring the stability of electrical power grids is critical to maintaining reliable energy distribution and mitigating the risk of blackouts, which can lead to substantial economic and safety challenges. Traditional stability prediction methods often rely on physical models constrained by inherent assumptions and limitations, making them less effective in addressing the dynamic and nonlinear characteristics of modern power grids. This study introduces a novel stacked machine learning approach to predict power grid stability by integrating ten distinct classifiers and corresponding regression models. Each model is rigorously evaluated for its predictive performance, contributing to a robust ensemble design that leverages the strengths of individual techniques. Experimental results demonstrate that the proposed stacked framework significantly outperforms conventional methods, achieving a remarkable 99% accuracy rate during validation. This high level of accuracy underscores the potential of machine learning ensembles in capturing complex patterns and interactions within power grid data. By employing advanced algorithms, this framework not only enhances prediction reliability but also contributes to proactive grid management strategies, ultimately fostering resilient and sustainable energy systems. The study highlights the practical implications of adopting machine learning-driven solutions to address critical challenges in power grid stability.

Accurate brain tumors segmentation is essential for precise diagnosis, planning, treatment, and monitoring of the tumor. However, the variations in tumor size, shape, and location, automating this process can be challenging. Segmentation using manual method is both time-taking and can produce inconsistent results. A promising solution to automated brain tumor segmentation is through deep learning approach that use convolutional neural networks. The models that have used U-Net shown considerable promise in enhancing the precision and effectiveness of automated Brain tumor segmentation. Still, there is a scope for improvement in accurate segmentation of brain MRI images. The segmentation of Brain MRI Images greatly helps in earlier diagnosis of Brain tumor effectively. The proposed model using ResNet, ResNext, and Inception as a backbone of U-Net with BraTs 2020 dataset could potentially reduce the workload of radiologists and provide more consistent results across experts. Future research can explore the use of other architectures or hybrid models to enhance brain tumor segmentation performance.

Nowadays, wine is being used extensively, and the industry is rapidly growing. Many wine industries are promoting their products to compete with the wine sector because it generates more revenue than other industries. Every wine industry aspires to make the best possible wines. Industries have started to discover the ideal combination of elements to balance wine quality by using historical and current data. Because of how well machine learning approaches work at revealing insights from both past and current data to improve decision-making in a variety of real-world issues and applications, such as those in the medical, educational, industrial, etc. To produce high-quality wine that will compete in the market, this study investigates how machine learning algorithms can assess data on wine in order to form inferences that will aid decision-makers in selecting the proper combination and amount of ingredients. This research work proposes a new hybrid Satin Bower Bird Optimization and Support Vector Machine (SBO-SVM) for wine quality prediction. Initially the collected data set is balanced by SMOTE ENN and utilize SVM for prediction. The performance of SVM is improved by SBO meta heuristic algorithm. The Experimental result shows that the hybrid SBO-SVM provide 95.8% accuracy which is higher than SVM without meta heuristic optimization.

Agriculture serves as a backbone for human life as well as for the economic development of the country. Major loss in agriculture is mainly due the disease which occurs on the plants. Crop diseases pose a significant threat to global food security, affecting the yield and quality of agricultural produce. Early detection of plant diseases is crucial for implementing timely interventions and minimizing crop losses. This research analyses the pretrained model to automate the detection of diseases in pepper and potato plants based on leaf images. This research analyzes the images of healthy and diseased pepper and potato leaves. The popular pre-trained CNN models VGG-16, Inception-v3 and Resnet are compared and finds the best algorithm for pepper and potato leaf disease prediction. According to experimental analysis, the Resnet-50 has obtained highest accuracy (100%), followed by VGG-16 (99 % ) and Inception-v3 (96 % ).

With the proliferation of the internet, big data continues to grow exponentially, and video has become the largest source. Video big data intro-duces many technological challenges, including compression, storage, trans-mission, analysis, and recognition. The increase in the number of multimedia resources has brought an urgent need to develop intelligent methods to organize and process them. The integration between Semantic link Networks and multimedia resources provides a new prospect for organizing them with their semantics. The tags and surrounding texts of multimedia resources are used to measure their semantic association. Two evaluation methods including clustering and retrieval are performed to measure the semantic relatedness between images accurately and robustly. A Fuzzy Rule-Based Model for Semantic Content Extraction is designed which performs classification with fuzzy rules. The features extracted are trained with the neural network where each network contains several layers among them each layer of neurons is dedicated to measuring the weight towards different semantic events. Each neuron measures its weight according to different features like shape, size, direction, speed, and other features. The object is identified by subtracting the background features and trained to detect based on the features like size, shape, and direction. The weight measurement is performed according to the fuzzy rules and based on the weight measures. These frameworks enhance the video analytics feature and help in video surveillance systems with better accuracy and precision.

Sentiment analysis has gained a lot of attention from researchers in the last year because it has been widely applied to a variety of application domains such as business, government, education, sports, tourism, biomedicine, and telecommunication services. Sentiment analysis is an automated computational method for studying or evaluating sentiments, feelings, and emotions expressed as comments, feedbacks, or critiques. The sentiment analysis process can be automated using machine learning techniques, which analyses text patterns faster. The supervised machine learning technique is the most used mechanism for sentiment analysis. The proposed work discusses the flow of sentiment analysis process and investigates the common supervised machine learning techniques such as multinomial naive bayes, Bernoulli naive bayes, logistic regression, support vector machine, random forest, K-nearest neighbor, decision tree, and deep learning techniques such as Long Short-Term Memory and Convolution Neural Network. The work examines such learning methods using standard data set and the experimental results of sentiment analysis demonstrate the performance of various classifiers taken in terms of the precision, recall, F1-score, RoC-Curve, accuracy, running time and k fold cross validation and helps in appreciating the novelty of the several deep learning techniques and also giving the user an overview of choosing the right technique for their application.