Volume 9, Issue 4

Aquaculture is one of the most important industries worldwide, and most marine products are produced by aquaculture. On the other hand, the aquaculture farmers are faced on the challenge of damage to marine products due to abnormal seawater temperatures. Research on seawater temperature prediction have been conducted, but many of them require a large amount of training data. Collecting seawater temperature data is not easy, and it takes an enormous time to introduce in new aquaculture farms. Therefore, the purpose of this study is to predict seawater temperature even with a small amount of training data for about one year. In this paper, we propose a seawater temperature prediction model using transition learning. The proposed model also attempts to improve the prediction accuracy by considering the difference in water depth between observation points. The results of the evaluation experiment showed that the prediction accuracy can be improved by transfer learning when learning with a small amount of data. In addition, we also confirmed that adding water depth values to the input layer may not lead to improved prediction accuracy for transfer learning.

This paper presents a method for automatic 3D segmentation of brain tumors in MRI using optical scanning holography. Automatic segmentation of tumors from 2D slices (coronal, sagittal and axial) enables efficient 3D reconstruction of the region of interest, eliminating the human errors of manual methods. The method uses enhanced optical scanning holography with a cylindrical lens, scanning line by line, and displays MRI images via a spatial light modulator. The outgoing phase component of the scanned data, collected digitally, reliably indicates the position of the tumor.The tumor position is fed into an active contour model (ACM), which speeds up segmentation of the seeding region. The tumor is then reconstructed in 3D from the segmented regions in each slice, enabling tumor volume to be calculated and cancer progression to be estimated. Experiments carried out on patient MRI datasets show satisfactory results. The proposed approach can be integrated into a computer-aided diagnosis (CAD) system, helping doctors to localize the tumor, estimate its volume and provide 3D information to improve treatment techniques such as radiosurgery, stereotactic surgery or chemotherapy administration. In short, this method offers a precise and reliable solution for the segmentation and 3D reconstruction of brain tumors, facilitating diagnosis and treatment.

RetePAIoT of Emilia-Romagna region is an IoT Public Network, financed by Emilia-Romagna Region and developed by Lepida Scpa, where citizens, private companies and Public Administrations can integrate free of charge their own sensors of any type and anywhere in the region. The main objective of the project is to provide a facility to implement the IoT paradigm, receiving data potentially from thousands of different sensors from the territory and to make them available to their owners and, in aggregate or anonymous form, to Public Administrations for their institutional purposes. In this context the interpretation of payloads sent by sensors, i.e. the extraction of the values measured by sensors, as well sharing them with all authorized subjects. are fundamental aspects that present a significant complexity due to the variety and unplannable context of the project. This paper illustrates the model and the architecture of a solution for the automatic extraction of values potentially coming from thousands of different sensors, which therefore requires a high level of flexibility, robustness and scalability as well as different methods for sharing them with third parties, depending on purposes and technical level required.

In healthcare, accurate communication is critical, and medical coding, especially coding using the ICD (International Classification of Diseases) standards, plays a vital role in achieving this accuracy. Traditionally, ICD coding has been a time-consuming manual process performed by trained professionals, involving the assignment of codes to patient records, such as doctor’s notes. In this paper, we present an automated ICD coding approach using deep learning models and demonstrate the feasibility and effectiveness of the approach across subsets of ICD codes. The proposed method employs a fine-grained approach that individually predicts the appropriate medical code for each diagnosis. In order to utilize sufficient evidence to enhance the classification capabilities of our deep leaning models, we integrate GPT-4 to extract semantically related sentences for each diagnosis from doctor’s notes. Furthermore, we introduce a hierarchical classifier to handle the large label space and complex classification inherent in the ICD coding task. This hierarchical approach decomposes the ICD coding task into smaller, more manageable subclassification tasks, thereby improving tractability and addressing the challenges posed by the high number of unique labels associated with ICD coding.