Style-Preserving Prompt Optimization for Game Item Images via Image-Based Prompt Extraction and Genetic Algorithms
이미지 기반 프롬프트 추출과 유전 알고리즘을 이용한 게임 아이템 이미지 스타일 유지형 프롬프트 최적화
This paper proposes a prompt–optimization framework for generating style–consistent game images using Stable Diffusion XL. Given a reference game item image, the system first extracts an initial prompt using a vision–language captioner and a domain–specific prompt bank. The extracted prompt is converted into a list of noun-like elements, and a genetic algorithm searches for compact combinations of these elements under dual gating based on SSIM and CLIP scores. The best combinations are treated as a “style template” that can reproduce the reference image with high structural and semantic similarity. We then investigate whether this template can be reused when the main object is changed while preserving the original visual style. Experiments on fantasy-style item images show that the framework reconstructs reference images using only 8–9 automatically discovered prompt elements, and that changing the main object token together with associated detail elements yields image sets that share a consistent visual style. In contrast, naïvely replacing only the main object token often produces visually ambiguous or stylistically inconsistent images. These results demonstrate that combining automatic prompt extraction from images with evolutionary optimization provides a concrete example of style–preserving prompt design for game item image generation.
목차
ABSTRACT 1. Introduction 2. Related Works 2.1 Game Images and Stylistic Consistency 2.2 Text-to-Image Generation and Image-Based Prompt Extraction 2.3 Automatic Prompt Optimization and Evolution-Based Search 3. Proposed System 3.1 Overview 3.2 Image-Based Automatic Prompt Extraction 3.3 Prompt Element List and Genetic Algorithm 3.4 Style Template and Object Replacement 4. Experiments and Results 4.1 Experimental Setup 4.2 Reference Image Reproduction Results 4.3 Object Replacement Results 5. Conclusion and Future Work Acknowledgement References
키워드
Game Item ImageStable Diffusion XLAutomatic Prompt ExtractionPrompt OptimizationGenetic AlgorithmCLIPSSIMStyle Consistency
저자
HyungJin Song [ 송형진 | Department of Computer Engineering, Hongik University ]
Jun Park [ 박준 | Department of Computer Engineering, Hongik University ]
Corresponding Author
1. 게임산업을 활성화 하고,
2. 게임기술과 기술 인력을 양산할 수 있도록 교육기관의 교과과정을 개발하고,
3. 관련기술에 대한 연구발표회, 강연회, 강습회 등을 개최하며,
4. 학회지, 논문지 및 관련 문헌을 발간하고,
5. 게임 기술 개발을 위한 국제화, 표준화 등을 지원하고,
6. 산.학.연.관이 협동할 수 있는 국제적 학술교류 및 협력을 지원하고,
7. 회원 상호간의 공동 이익과 친목을 증진시킨다.
간행물
간행물명
컴퓨터게임및콘텐츠논문지(구 한국컴퓨터게임학회논문지) [Journal of Computer Games and Contents]
간기
월간
pISSN
3091-7409
eISSN
3092-3638
수록기간
2002~2026
등재여부
KCI 등재
십진분류
KDC 691DDC 793
이 권호 내 다른 논문 / 컴퓨터게임및콘텐츠논문지(구 한국컴퓨터게임학회논문지) 제38권 제8호