Seung-Soo Shin

Ting Chai, Sung-Hwa Han, Seung-Soo Shin, Member, KIICE. Journal of information and communication convergence engineering 2026;24:49-58. https://doi.org/10.56977/jicce.2026.24.1.49

Recent advances in deep learning have facilitated widespread deployment of posture recognition system. However, the scarcity of accessible human posture datasets presents a significant challenge to achieving high model precision and robustness. To address this challenge, we introduce a federated learning (FL) framework that enables training of posture recognition models across diverse and privacy-sensitive datasets. Our approach integrates FL strategies with fundamental deep learning architectures for human pose estimation and action recognition. To identify the optimal base model, we evaluate five widely adopted architectures. Using a posture recognition dataset, our systematically compare conventional single-machine training models with FL-enhanced models across key metrics: precision, recall, F1-score, and accuracy. Results demonstrate that our FeLM achieves superior generalization while preserving data privacy, highlighting the efficacy of FL in posture recognition tasks. This underscores the benefit of FL in improving the generalization of a model without compromising data privacy.

Mang Lu, Sung-Hwa Han, and Seung-Soo Shin. Journal of information and communication convergence engineering 2025;23:113-20. https://doi.org/10.56977/jicce.2025.23.2.113

With the rapid advancement of network technology, network security issues have become increasingly prominent, particularly the frequent occurrence of malicious behaviors that pose significant threats to both individuals and organizations. However, attackers are now more easily bypassing traditional network intrusion detection systems. The detection of such malicious behavior is crucial, regarding data security and impacts the system's stability and credibility. This study explored the application of deep learning for detecting such actions and analyzed its advantages over traditional methods. For comparative analysis, we employed a variety of machine learning algorithms, including CNNs, random forests, and support vector machines,. Additionally, we proposed a novel model that combines a CNN with long short-term memory networks. We experimentally verified the algorithm's effectiveness in practical applications.

Efficient communication in the Internet of Things (IoT) is essential for enabling smart applications. While NFC excels in near-field device interaction, its limited communication range hinders LoRa’s long-range communication due to its low data throughput. Together with NFC and LoRa technologies, Raspberry Pi is used as a microcontroller (MCU) in this paper to look into how to make near-field and long-distance communication work better together and fix the issue of an imbalance between communication range and energy consumption in the IoT system. By optimizing the communication algorithm and parameter tuning, the power consumption of the system is significantly reduced, and the communication range and data throughput are improved. Our research gives you the technical information you need to make an IoT communication system that works well, uses little power, and has a wide coverage area. This kind of system is good for situations where you need to collect data from a close distance and keep an eye on things from afar. This makes the system more power-efficient and better at communicating, which also makes it easier for users to manage data. It is suitable for a wide range of application scenarios, such as warehousing, healthcare, agriculture, and smart cities.

The exponential growth of the Internet of Things has resulted in a substantial surge in the volume of data produced by numerous sensors and microcontroller devices. The objective of this study is to create a comprehensive data-management framework using blockchain and InterPlanetary File System with the aim of enhancing the security, dependability, and decentralized processing of large-scale data storage. This study utilizes a hyperledger blockchain and an interstellar file system to transmit data via ZigBee and Wi-Fi networks. In addition, edge devices are employed for pre-processing to guarantee the security of data storage and exchange. The performance of the model in terms of data management is verified and analyzed using simulations. The results demonstrate a significant enhancement in both the data-storage access speed and overall system efficiency owing to the implementation of the model. This study aims to offer efficient methods for managing large amounts of data in the Internet of Things context.