BROWSE
Detail
Conference Paper
이미지 개인화 기술의 정면 뷰 편향 생성 완화 연구
Cited - time in 
Share 
Share
- Issue Date
- 2026-06
- Citation
- 대한전자공학회 학술 대회 (하계) 2026, pp.1-3
- Publisher
- 대한전자공학회
- Language
- Korean
- Type
- Conference Paper
- Abstract
- Image personalization enables users to generate images incorporating user-defined subjects. However, the generated results are often biased toward front views. In this paper, we propose a method to mitigate the frontal-view bias problem in image personalization. First, we construct multiview image-text pair dataset from user-defined subject images. Then we train a personalized model using view images with text prompts that explicitly include viewpoint information. Experimental results demonstrate that the proposed method effectively mitigates the problem.
- KSP Keywords
- Image-text, Personalized model, multi-view images, user-defined