Conference
Order-Conditioned Message Passing over Shape Codes for Amodal Segmentation
Vungsovanreach Kong, Saravit Soeng, Rotanakkosal Chhun, Tae-Kyung Kim
BIGDAS 2026, 14th International Conference on Big Data Applications and Services, Aug. 2026
Research team
SOENG Saravit
Researcher
AI Convergence Lab, Chungbuk National University
- Computer Vision
- LLM
- VLM
- Applied AI
About
Researcher at AI Convergence Lab (AICLab), Chungbuk National University, with over 9 years in software development and 5+ years specializing in machine learning, deep learning, and applied AI. His research focuses on multimodal AI systems, real-time abnormal action detection, computer vision, and vision-language models (VLM).
Education
PhD
Chungbuk National University, South Korea
Dissertation: "A Multimodal Framework for Real-time Abnormal Human Action Detection Using Spatio-Temporal Modeling and Vision-Language Reasoning"
Ph.D. in Big Data
(Mar 2022 – Feb 2026)Chungbuk National University, South Korea
Dissertation: "A Multimodal Framework for Real-time Abnormal Human Action Detection Using Spatio-Temporal Modeling and Vision-Language Reasoning"
MS
Chungbuk National University, South Korea
Thesis: "An Implementation of IoT End-to-end Data Pipeline for Industrial Big Data Platform"
M.S. in Big Data
(Mar 2020 – Feb 2022)Chungbuk National University, South Korea
Thesis: "An Implementation of IoT End-to-end Data Pipeline for Industrial Big Data Platform"
BS
Royal University of Phnom Penh, Cambodia
B.S. in Computer Science
(Nov 2013 – May 2017)Royal University of Phnom Penh, Cambodia
Contact
AI Convergence Lab, Chungbuk National University, Cheongju, South Korea
Recent publications
View Conference
Conference
A Verification Agent Harness for Enhancing Factual Reliability in Enterprise RAG Systems
Saravit Soeng, Vungsovanreach Kong, Sokheang Chan, Tae-Kyung Kim
BIGDAS 2026, 14th International Conference on Big Data Applications and Services, Aug. 2026
Conference
Multi-Agent Scientific Hypothesis Generation & Verification
Naro Vorn, Ratanaktepi Chhor, Saravit Soeng, Tae-Kyung Kim
BIGDAS 2026, 14th International Conference on Big Data Applications and Services, Aug. 2026
Research Focus
Multimodal Action Detection
Spatio-temporal modeling and vision-language reasoning for real-time abnormal human action detection (fall detection, fight identification, safety compliance)Computer Vision
Object detection, pose estimation, image classification, and intelligent video surveillance systemsVision-Language Models (VLM)
Multimodal AI combining visual and language understanding for scene reasoning and document intelligenceLarge Language Models (LLM)
Applied LLM systems including OCR-based document extraction, NLP pipelines, and prompt-driven solutionsIoT & Big Data Infrastructure
End-to-end IoT data pipelines for industrial sensor data collection, processing, storage, and monitoring using Spark and KafkaApplied AI Systems
Address verification with geocoding and deep learning, blockchain-based voting systems, and real-world AI deploymentProfessional Experience
- 2026 – PresentResearcher, AI Convergence Lab, Chungbuk National University
- 2022 – 2026Research Assistant (PhD), Big Data Lab, Chungbuk National University
- 2020 – 2022Research Assistant (MSc), Big Data Lab, Chungbuk National University
- 2017 – 2020Instructor & Developer, Korea Software HRD Center, Cambodia
2026
Conference
Order-Conditioned Message Passing over Shape Codes for Amodal Segmentation
Vungsovanreach Kong, Saravit Soeng, Rotanakkosal Chhun, Tae-Kyung Kim
BIGDAS 2026, 14th International Conference on Big Data Applications and Services, Aug. 2026
Conference
A Verification Agent Harness for Enhancing Factual Reliability in Enterprise RAG Systems
Saravit Soeng, Vungsovanreach Kong, Sokheang Chan, Tae-Kyung Kim
BIGDAS 2026, 14th International Conference on Big Data Applications and Services, Aug. 2026
Conference
Multi-Agent Scientific Hypothesis Generation & Verification
Naro Vorn, Ratanaktepi Chhor, Saravit Soeng, Tae-Kyung Kim
BIGDAS 2026, 14th International Conference on Big Data Applications and Services, Aug. 2026
2025
SCIE
PIFR: A novel approach for analyzing pose angle-based human activity to automate fall detection in videos
Vungsovanreach Kong, Saravit Soeng, Munirot Thon, Wan-Sup Cho, Anand Nayyar, Tae-Kyung Kim
PLOS ONE, Jun. 2025
2024
SCOPUS
A Simple and Efficient Approach for Extracting Object Hierarchy in Image Data
Saravit Soeng, Vungsovanreach Kong, Munirot Thon, Wan-Sup Cho, Tae-Kyung Kim
International Journal of Advanced Computer Science and Applications, Vol. 15, No. 8, Aug. 2024