Kyungjune Baek
연세대학교 컴퓨터학과 · 컴퓨터과학
Kyungjune Baek 교수의 연구실은 딥러닝 기반 이미지 생성 및 변환 기술에 초점을 맞추고 있으며, 특히 레이블이 부족한 환경에서도 효과적으로 학습할 수 있는 비지도 학습 및 자기지도 학습 기반 기법을 개발하고 있습니다. 대표적으로 이미지 간 변환, 얼굴 이미지의 동시에 생성과 속성 편집, 그리고 저샷 러닝 환경에서의 GAN 전이 학습 등 고해상도 이미지 생성과 보안·일관성 문제 해결을 동시에 고려한 연구를 진행하고 있습니다. 특히, 데이터의 비공개성과 저질 레이블 문제에 대응하기 위한 혁신적인 데이터 생성 및 정제 기법 개발도 핵심 과제입니다.
표시된 성과는 수집된 데이터 기준으로 산출되며, 일부 차이가 있을 수 있습니다.
Every recent image-to-image translation model inherently requires either image-level (i.e. input-output pairs) or set-level (i.e. domain labels) supervision. However, even set-level supervision can be a severe bottleneck for data collection in practice. In this paper, we tackle image-to-image translation in a fully unsupervised setting, i.e., neither paired images nor domain labels. To this end, we propose a truly unsupervised image-to-image translation model (TUNIT) that simultaneously learns t
Existing co-localization techniques significantly lose performance over weakly or fully supervised methods in accuracy and inference time. In this paper, we overcome common drawbacks of co-localization techniques by utilizing self-supervised learning approach. The major technical contributions of the proposed method are two-fold. 1) We devise a new geometric transformation, namely point symmetric transformation and utilize its parameters as an artificial label for self-supervised learning. This
Transfer learning for GANs successfully improves generation performance under low-shot regimes. However, existing studies show that the pretrained model using a single benchmark dataset is not generalized to various target datasets. More importantly, the pretrained model can be vulnerable to copyright or privacy risks as membership inference attack advances. To resolve both issues, we propose an effective and unbiased data synthesizer, namely Primitives - PS, inspired by the generic characterist
The remarkable performance of deep neural networks heavily rely on large-scale datasets with high-quality annotations. Since the data collection process such as web crawling naturally involves unreliable supervision (i.e., noisy label), handling samples with noisy labels has been actively studied. Existing methods in learning with noisy labels (LNL) 1) develop the sampling strategy for filtering out the noisy labels or 2) devise the robust loss function against noisy labels. As a result of these
We propose a novel framework for simultaneously generating and manipulating the face images with desired attributes. While the state-of-the-art attribute editing technique has achieved the impressive performance for creating realistic attribute effects, they only address the image editing problem, using the input image as the condition of model. Recently, several studies attempt to tackle both novel face generation and attribute editing problem using a single solution. However, their image quali
Transfer learning for GANs successfully improves generation performance under low-shot regimes. However, existing studies show that the pretrained model using a single benchmark dataset is not generalized to various target datasets. More importantly, the pretrained model can be vulnerable to copyright or privacy risks as membership inference attack advances. To resolve both issues, we propose an effective and unbiased data synthesizer, namely Primitives-PS, inspired by the generic characteristic
Every recent image-to-image translation model inherently requires either image-level (i.e. input-output pairs) or set-level (i.e. domain labels) supervision. However, even set-level supervision can be a severe bottleneck for data collection in practice. In this paper, we tackle image-to-image translation in a fully unsupervised setting, i.e., neither paired images nor domain labels. To this end, we propose a truly unsupervised image-to-image translation model (TUNIT) that simultaneously learns t
