Search

This paper aims to address the detection of COVID-19 by developing an accurate and efficient diagnostic system using chest X-ray images. The research utilizes open-source Kaggle data comprising four categories: COVID-19, Lung-Opacity, Normal, and Viral Pneumonia. The proposed system employs convolutional neural networks (CNNs), including VGG19, RNN-LSTM, and inceptionv3. Results vary among the methodologies, with VGG19 achieving 26% accuracy, RNN-LSTM attaining 25% accuracy (28% with preprocessing), and inceptionv3 with histogram equalization achieving 83% accuracy. A CNN designed from scratch demonstrates the highest performance, with an accuracy of 93% (96% with histogram equalization). The findings emphasize the potential of AI techniques in enhancing disease diagnosis, particularly in distinguishing COVID-19 from other conditions, thereby facilitating timely and effective interventions. [ABSTRACT FROM AUTHOR]

Examining facial expressions is a crucial aspect that garners attention due to its significance in divulging emotional states. This study delves into employing a robust deep learning method for automatically analyzing and identifying facial emotions in images. The chosen technique revolves around the convolutional neural network (CNN) algorithm. A dataset containing images of individuals, each exhibiting distinct facial expressions, was curated. The emotions in these images were categorized into seven groups (angry, disgust, fear, happy, neutral, sad, surprise) based on the depicted emotional states. The approach comprises four primary steps: preprocessing the input facial images, utilizing image adjustments and data augmentation, employing the Viola and Jones technique for face detection and landmark localization, creating a numerical feature vector from the registered image for feature extraction, and inputting the extracted features into the CNN for classification. The proposed CNN model underwent application to classify facial emotions within the image dataset. Additionally, a pretrained VGG-16 model was incorporated into the classification process for facial images. A comparative analysis between the proposed CNN approach and the pretrained VGG-16 model revealed that the latter outperformed the former in terms of accuracy rate and loss function values when determining individuals’ facial expressions. [ABSTRACT FROM AUTHOR]

Object recognition and detection is an area of study, within intelligence and computer vision. It finds applications in fields such as surveillance, detailed activity analysis, robotics and object tracking. The primary focus of research papers in this domain revolves around enhancing the precision of object identification and detection regardless of whether the objects are located indoors or outdoors. To address this challenge, a new approach involving the utilization of SIFT features for information extraction has been proposed. Our approach encompasses two components; the implementation of dimensionality reduction through principal component analysis (PCA) to eliminate redundancies; secondly the incorporation of feature vector encoding using fisher encoding techniques. The RGB-D dataset employed contains 300 objects across scenarios with emphasis on colored aspects rather than depth. The SIFT features are categorized using a support vector machine (SVM) into 7 classes. When compared to using SIFT features integrating them with encoding methods notably enhances recall, precision and F1-score by than 30% through fisher encoding and PCA techniques. The study concludes with an evaluation based on n-cross validation methodology along, with detailed experimental results. [ABSTRACT FROM AUTHOR]