YOLO ALGORITHM USING IN DECISION SUPPORT SYSTEMS BASED ON THE USE OF NEURAL NETWORKS

In recent years, computer technologies, particularly computer vision technologies, have played an increasingly important role in medical diagnostics, including otoscopy—the process of visualizing the external auditory canal and tympanic membrane to detect ear pathologies. However, otoscopic images have several characteristics that complicate automatic classification, such as variations in lighting, orientation, and the presence of artifacts. These factors necessitate the development of reliable deep learning models for accurate disease classification. This paper examines modern approaches to address this issue. It discusses the use of convolutional neural networks (CNNs) for object detection in medical images, specifically photographs and videos obtained through otoscopic examinations, to facilitate diagnostic procedures. An analysis of the OTONet architecture, specifically designed for classifying otoscopic images, as well as research utilizing quantum machine learning (QML), is conducted. The paper explores a hybrid quantumclassical CNN architecture that employs both quantum and classical convolutional layers for extracting hidden features from images. The use of deep learning architectures, such as CNNs, integrated with the YOLO algorithm, is also discussed as a promising approach to automating diagnostics using otoscopic images. The training process of the model is described, which involves function approximation based on examples and the use of one-hot encoding for convenient handling of categories. The network predicts objects by dividing the image into a grid of cells and assigning bounding boxes to each object, taking into account their position, size, and confidence level. This approach allows for the effective resolution of complex classification and object detection tasks with high accuracy. This line of research holds significant potential for further development in the medical field, providing effective solutions for medical image processing and improving diagnostic quality.