KJCKorea
Journal Central

Blood is composed of white blood cells (WBC), red blood cells (RBC), and platelets, components closely related to human health. They respectively contribute to the immune system, the transfer of oxygen to body organs, and bleeding prevention. In the case of WBC, when abnormalities occur in the body, such as infections and allergic reactions, the relevant type of WBC in the blood increases to respond. Therefore, it is possible to identify the type of WBC in the case of abnormalities in the body through the application of deep learning-based object detection algorithms and classification methods. This information can be used to predict the health status of patients easily. Therefore, this paper proposes a method for extracting the region of interest (RoI) for WBC and classifying WBC types based on blood microscope images using you-only-look-once (YOLO) and feature ensemble techniques. We select RoI extraction models through a comparative analysis of WBC detection performance for YOLO V5, V8, V9 and YOLO-Neural Architecture Search (YOLO-NAS), and demonstrate an improvement in WBC type classification performance through comparative analysis and a Top-3 model feature ensemble based on general convolutional neural network (CNN) models, such as ResNet and EfficientNet. Compared to the approximately 98% performance of the Top-3 models based on the F1-Score, it achieved an improved performance of approximately 99%.

In this paper, we propose MetaPics, a web-based image management system that integrates object recognition and metadata extraction. The surge in image data and the resulting increase in management complexity are major challenges, and object recognition and image metadata utilization are emerging as key elements of image data search and management. MetaPics based on object recognition algorithms has improved precision, average precision (mAP), and recall, which are object detection performance within an image. MetaPics is a system with scalability and user-centered design that can be applied to various applications in the field of artificial intelligence and digital content management. It solves the complexity of image data management and provides a user-centered, intuitive user experience. MetaPics enables efficient and systematic management of image data through accurate object recognition and image metadata utilization.

This paper proposes the use of anomaly detection using LSTM AutoEncoder to verify the possibility of anomaly detection function through AI in the control environment of the interlocking weapon system of a naval combat system.. Performance data such as combat system logs and metric data were collected and time-series preprocessed with the ELK Stack. The LSTM AutoEncoder model, which uses the LSTM network-based Eocder to compress and dimensionally reduce data, and the Decoder to restore the input data to a similar form, was trained using only normal environmental data. Afterwards, the performance was evaluated using test data generated by simulating normal and abnormal situations, and a high score of Accuracy 0.99, Precision 0.97, Recall 0.87, and F1-Score 0.92 was output. This study confirmed the applicability of the model generated through machine learning in the detection of anomalies in the control environment of the interlocking weapon system of a naval combat system.