Research Background
The Intelligent Media Laboratory (IMLab) focuses on building practical AI systems that can understand complex real-world environments and support decision-making in safety-critical and resource-constrained settings. Our research spans anomaly understanding for smart surveillance, energy and smart-grid intelligence, and data-driven platforms for societal and industrial challenges. We develop machine learning and deep learning methods that are robust, explainable, and deployable,bridging algorithm design with real-world applications.
Research Projects Overview
Our current and recent projects can be broadly grouped into three directions:
1) Cross-View Adaptive Representation Learning Based Early Prediction Method for Abnormal Behaviors in Multi-space Environments
Cross-View Adaptive Representation Learning Based Early Prediction Method for Abnormal Behaviors in Multi-space Environments (03/2026 – 02/2029)
This project develops a unified AI framework for early prediction of abnormal behaviors across multiple interconnected spaces in smart-city environments.
2) Anomaly Understanding and Connected Vision for Safety
We study video-based behavior understanding to detect abnormal events and prevent accidents in connected and intelligent environments.
Anomaly Behavior Recognition for Accident Prevention in the Connected Vision Environment (03/2023 – 02/2026)
This project develops AI techniques to recognize risky and anomalous behaviors in connected-vision environments.
Multi-view Video Data Analysis Technology for Smart City based Intelligent Surveillance System (06/2019 – 02/2022)
We develop multi-view analysis methods for smart-city surveillance.
Multi-Viewed Video Data Processing and Analysis Techniques for Intelligent Video Surveillance System (06/2016 – 05/2019)
This project focuses on scalable video processing pipelines and learning-based analysis for intelligent surveillance applications.
3) Energy AI and Smart Network Optimization
We develop data-driven methods for energy systems, focusing on interpretability, forecasting, optimization, and platform technologies for next-generation energy networks. Our research supports efficient energy usage analysis, network participation, and AI-enabled policy simulation. Open Platform Technology for Autonomous Participation of Energy Network Components (06/2019 – 12/2023) We study open platform technologies that enable energy network components to participate autonomously, supporting flexible and scalable energy ecosystem operation. Development of Ensemble Methods-Based XAI Energy Platform for Effective Energy Consumption Pattern and Factor Analysis (06/2019 – 12/2023) This project develops explainable AI (XAI) approaches for analyzing energy consumption patterns and identifying influential factors for improved energy management.
4) Data Platforms for Societal and Industrial Impact
We build AI and data platforms that connect learning, simulation, and data-cloud infrastructure to address large-scale societal and industrial problems, including depopulated regions and materials informatics.
Development and Demonstration of AI Policy Simulation Platform Technology to Solve Social Problems in Depopulated Areas Based on XOps (04/2024 – 12/2027)
This project aims to build an AI policy simulation platform that supports decision-making for depopulated regions by integrating scalable operational workflows and data-driven policy evaluation.
Data HUB for Solid Electrolyte Materials Based on SyncroLab Data Cloud (07/2024 – 12/2025)
We develop a data-hub platform for solid electrolyte materials, enabling structured data collection, retrieval, and analysis through a cloud-based pipeline to accelerate materials research and discovery.
Development of AI-Convergence Technologies for Smart City Industry Productivity Innovation (04/2019 – 12/2021)
This project develops AI convergence technologies to improve smart-city industry productivity through practical AI solutions and data-driven services.
Development of Integrated Cross-Model Data Processing Platform Supporting a Unified Analysis of Various Big Data Models (04/2019 – 12/2021)
We build an integrated platform to support unified analysis across heterogeneous big data models, enabling scalable processing and cross-model insights.
National Program of Excellence in Software (10/2015 – 12/2021)
This program supports advanced software education and research, strengthening capabilities in AI-driven software systems and applied R&D.
Keywords
Anomaly Recognition · Video Understanding · Smart Surveillance · Energy AI · Explainable AI · Optimization · Data Platforms · Policy Simulation · Materials Data Cloud · Smart Cities