중앙대학교 첨단영상대학원

구성원

전임

VILab에서는 이미지와 영상에 기반한 최신의 딥러닝 모델과 학습 방법을 연구하고 있습니다. 또한, 웹툰, 애니메이션, 문화유산 등 다양한 콘텐츠 분야에 활용할 수 있는 다양한 인공지능 기반의 시스템을 설계하고 개발하는 연구를 진행합니다. 대표적인 연구 주제로 Domain Adaptation, Active Learning, Deepfake Detection, AI-based Content Platform 등이 있습니다.

2021.11 - Outstanding Paper, Korean Artificial Intelligence Association Conference (2021 Fall)
2020.01 - Excellence Award on Research Activities, Samsung SDS
2017.02, 2018.02 - Best Paper Award, 30th Workshop on Image Processing and Image Understanding (IPIU2018, IPIU2017)
2013.08 - Qualcomm Fellowship Award 2013

FingerprintNet: Synthesized Fingerprints for Generated Image Detection, European Conference on Computer Vision (ECCV), 2022
FrePGAN: Robust Deepfake Detection Using Frequency-level Perturbations, AAAI Conference on Artificial Intelligence (AAAI), 2022
VaB-AL: Incorporating Class Imbalance and Difficulty with Variational Bayes for Active Learning, Conference on Computer Vision and Pattern Recognition (CVPR), 2021
Visual Domain Adaptation by Consensus-based Transfer to Intermediate Domain, AAAI Conference on Artificial Intelligence (AAAI), 2020
Context-aware Deep Feature Compression for High-speed Visual Tracking, Conference on Computer Vision and Pattern Recognition (CVPR), 2018
Attentional Correlation Filter Network for Adaptive Visual Tracking, Conference on Computer Vision and Pattern Recognition (CVPR), 2017