Optical coherence tomography (OCT) is widely used to detect and classify retinal diseases. However, OCT-image-based manual detection by ophthalmologists is prone to errors and subjectivity. Thus, various automation methods have been proposed; however, improvements in detection accuracy are required. Particularly, automated techniques using deep learning on OCT images are being developed to detect various retinal disorders at an early stage. Here, we propose a deep learning-based automatic method for detecting and classifying retinal diseases using OCT images. The diseases include age-related macular degeneration, branch retinal vein occlusion, central retinal vein occlusion, central serous chorioretinopathy, and diabetic macular edema. The proposed method comprises four main steps: three pretrained models, DenseNet-201, InceptionV3, and ResNet-50, are first modified according to the nature of the dataset, after which the features are extracted via transfer learning. The extracted features are improved, and the best features are selected using ant colony optimization. Finally, the best features are passed to the k-nearest neighbors and support vector machine algorithms for final classification. The proposed method, evaluated using OCT retinal images collected from Soonchunhyang University Bucheon Hospital, demonstrates an accuracy of 99.1% with the incorporation of ACO. Without ACO, the accuracy achieved is 97.4%. Furthermore, the proposed method exhibits state-of-the-art performance and outperforms existing techniques in terms of accuracy.

Accurate sleep posture monitoring plays a pivotal role in the diagnosis and treatment of a spectrum of sleep-related disorders, such as sleep apnea, restless leg syndrome, and rapid eye movement sleep behavior disorder. These disorders can significantly affect an individual’s overall well-being and quality of life. However, the complexity of the human body and various factors affecting sleep patterns pose substantial challenges. Thus, this study introduces a sleep position determination approach that utilizes both RGB and thermal cameras to acquire visual and thermal data. This dual-source data acquisition system enables a comprehensive analysis of body positioning during sleep, including potential discomfort assessment. The methodology for this endeavor was initiated by acquiring a dataset encompassing various sleep postures, which was achieved through the deployment of RGB and thermal cameras. This dataset includes video footage of five frequently adopted sleep positions: supine, left log, right log, prone left, and prone right, involving the participation of nine individuals. Furthermore, the dataset was collected under two conditions: with and without a blanket. The proposed method begins by normalizing the database to the video frames. Next, the fine-tuned MobileNet-V2 and Inception-V3 models were employed for feature extraction. The tree-seed algorithm was used to select optimal features from the extracted data, reduce dimensionality, and improve the classification performance. Subsequently, Parallel Standard Deviation Padding Max Value (PSPMV), was applied to combine the feature vectors from the RGB and thermal datasets to enhance the accuracy. The fused vectors are then classified using ensemble machine learning models. Our method achieved an accuracy of 97.8% and 98.4% with and without blanket, using subspace KNN. After applying PSPMV fusion to blanket features, the accuracy improved to 99.2%.

This study explores a model for detecting fall, non-fall and near-fall events as frequent experiences of near-falls are closely associated with a heightened risk of falls. Detecting near-falls can lead to more accurate predictions of falls. However, near-falls exhibit certain movement patterns similar to actual falls, making it challenging to distinguish between the two. We investigate the detection of fall-related activities, including falls, near-falls, and non-falls, by utilizing dynamic motion images derived from video clips. There are two primary approaches for classification: a vanilla convolutional neural network (CNN) model and a transfer learning approach that utilizes InceptionV3 and DenseNet201 models as feature extractors and train conventional machine learning classifiers, such as support vector machine (SVM), K-nearest neighborhood, decision tree, and random forest, and adaptive boosting models. The vanilla CNN model achieved a high accuracy of 97.89% compared to the transfer learning approach, which reached a maximum accuracy of 95.31 for binary classification of fall and non-fall events. On the other hand, the transfer learning approach, which integrated feature from InceptionV3 and DenseNet201 into machine learning classifiers, achieved an accuracy of up to 90.14% for the three-class classification of fall, non-fall, and near-fall events. This underscores the model’s robustness in detecting various fall-related activities, highlighting its potential for improving safety in at-risk populations.