ㅇ Real-Time Network Anomaly Detection

This study explores real-time network anomaly detection using advanced deep learning techniques, including Gated Graph Neural Networks (GGNN) and Long Short-Term Memory (LSTM) models. The research focuses on analyzing flow patterns of widely used network protocols such as HTTP and TELNET. By modeling protocol-level correlations and structural flow behaviors, the proposed approach identifies abnormal signals that deviate from normal operational patterns. This work contributes to the development of intelligent, AI-driven cybersecurity systems capable of detecting network anomalies at the protocol level.

ㅇ ECG-Based Advanced Personal Identification

Each individual possesses unique biological information that rarely overlaps with that of others. Based on this characteristic, numerous studies have explored user identification methods using biometric signals. However, dynamic biological signals are not always reproduced identically across different times and environments, making exact matching difficult. Although bio-information varies whenever it is measured, natural signals inherently contain individual-specific patterns. In this paper, ECG (electrocardiogram) signals are employed as biometric information to extract distinctive patterns within a defined threshold range. User identification is performed by detecting patterns that fall within this threshold. For analysis, ECG signals are recorded for 120 seconds and subsequently segmented into intervals ranging from 1–2 seconds to 5 seconds in order to examine the stability and individuality of the extracted patterns.

ㅇ Secure Energy Trade System with Blockchain

Transformative computing in the era of the Fourth Industrial Revolution collects signals and sequential data from sensing devices through AI-based wireless networks and integrates them into meaningful information. In industrial and residential environments, numerous electronic devices generate diverse energy consumption patterns. Without a smart energy management system to monitor energy flow, significant energy waste or even blackouts may occur.

By analyzing energy flow, peak and high-demand periods can be identified, enabling more efficient energy distribution. In this study, a smart energy management system establishes an adaptive control strategy based on analyzed consumption patterns, regulating energy usage accordingly. The system also reduces power to idle or unnecessary IoT devices and detects human behavior to automatically control device activation, thereby improving overall energy efficiency.

ㅇ Intelligent Virtual-World System (IVWS)

2008년, 포르토에서 아인슈타인의 이론에 빠져들었다. 유튜브를 보며 문득 이런 생각이 들었다. “이런 것을 공부하면 참 재미있지 않을까?” 그렇게 순수한 호기심으로, 이해하고 싶다는 마음, 보이지 않는 질서를 탐구하고 싶다는 충동. 당시에는 이 작은 관심이 훗날 Cyber-Digital-Contents, 즉 AR·VR·MR로 확장될 것이라고는 상상하지 못했다. 그저 재미있을 것 같아서 시작했을 뿐. 그러나 사유는 또 다른 사유를 낳고, 질문은 더 깊은 질문으로 이어졌다. 기술은 철학을 만나 확장되었고, 관심은 연구가 되었다. 돌이켜보면, 그때의 호기심이 오늘날 딥페이크 탐지와 디지털 미디어 보안 연구의 한 토대가 될 것이라고 누가 예상했겠냐. "연구란 결국 거대한 계획이 아니라, 작은 궁금증이 시간 속에서 자라나는 과정일뿐" 

In 2008 Porto, I became absorbed in Einstein’s theories. Watching lectures on YouTube, a simple thought crossed my mind: “Wouldn’t it be fascinating to study this?” What began was not a grand research agenda, but pure curiosity — a desire to understand, to explore the invisible order beneath phenomena. At that time, I could not have imagined that this small intellectual spark would later expand into what is now Cyber-Digital-Contents research, including AR, VR, and MR technologies. I started simply because it seemed interesting. Yet curiosity generates reflection, and reflection leads to deeper questions. Technology met philosophy, and interest gradually transformed into research. Looking back, who could have anticipated that this early curiosity would eventually contribute to my work in deepfake detection and digital media security? Perhaps research is not born from strategic intention, but from small questions that quietly grow over time.