Design and Implementation of Framework for Automatically Generating Game Map Data using DCGAN Learning Model

doi:10.22716/sckt.2023.11.3.030

All Issue

2023 Vol.11, Issue 3 Preview Page

Research Article

Design and Implementation of Framework for Automatically Generating Game Map Data using DCGAN Learning Model DCGAN 학습 모델을 이용한 게임 맵 데이터 자동 생성 프레임워크 설계 및 구현: HyeonWoo Nam¹
남현우¹; ¹Professor, Department of Computer Science, Dongduk Women’s Univ.

¹동덕여자대학교 컴퓨터학과 교수

30 September 2023. pp. 113-122

PDF

Abstract

코로나 팬데믹의 여파로 실내 활동 증가와 여성 게이머의 증가로 성장하는 모바일 게임 장르 중 캐주얼 게임과 퍼즐 게임의 비중이 높아지고 있다. 이에 따라 장르에 대한 개발 수요가 증가하고 있지만, 개발 과정에서 필수적인 다양한 게임 맵 데이터를 생성하기 위해서는 많은 시간이 소요된다. 본 논문에서는 DCGAN 모델을 이용하여 게임 맵 데이터를 자동으로 생성하기 위한 프레임워크를 설계하고 구현하였다. 제안된 프레임워크 구현에 적용된 기술로 인해 전체 개발 기간을 3분의 1로 단축할 수 있다는 만족스러운 결과를 얻었다. 또한, 자동 생성된 맵 데이터의 등급 분류 정확도 검증 기능을 적용하여 원본 데이터를 기반으로 생성된 맵 데이터의 진위 여부를 확인할 수 있었다. 개발된 프레임워크에 대한 기술이 고도화된다면 다른 장르의 모바일 게임 개발에도 적용될 수 있을 것으로 기대된다.

The proportion of casual games and puzzle games among the genres of mobile games that are growing due to the increase in indoor activities and the increase in female gamers under the influence of COVID-19 is increasing. Accordingly, the development demand for the genre is increasing, but it takes a lot of time to generate various game map data essential in the development process. In this paper, a framework for automatically generating game map data using DCGAN model was designed and implemented. Due to the technology applied to the proposed framework implementation, a satisfactory result was obtained that the entire development period could be reduced by a third. In addition, it was possible to verify the authenticity of the map data generated based on the original data by applying the map class classification accuracy verification function of the automatically generated map data. If the technology for the developed framework is advanced, it is expected that it can be applied to the development of other genres of mobile games.

Keywords

PGAN(Generative Adversarial Networks)

DCGAN(Deep Convolutional GAN)

MAP

FID(Frechet Inception Distance)

TSNE(T-distributed Stochastic Neighbor Embedding)

References

Newzoo, "2023 Global Games Market Report", 2023.
KOCCA, "2022 Korean Game White Paper", 2023.
https://newsinside.tistory.com/38 , 2022.
KOCCA, "글로벌 게임산업 트렌드(2023년 1+2월호)", 2023.
Y. R. Sim, H. Y. Kang, and J. H. Han, "Multi-resolution Perlin Noise Decomposition and Procedural Texture Synthesis by Example", In Conference on HCI Korea, Vol. 2016, No. 1, pp. 382-388, 2016.
J. W. Ko, B. P. Kyung, and D. Y. Lee, "Procedural Map Generator Algorithm for Variableness Game World", In Conference on Korea Game Society, pp. 159-162, 2017.
C. J. Hart, "Perlin noise pixel shaders", Proceedings of the ACM SIGGRAPH/EUROGRAPHICS workshop on Graphics hardware. pp. 87-94, 2001. 10.1145/383507.383531 11172120 PMC4096700
K. M. Jeong, "Implementation of a 3D Terrain Map Editor Based on Height Map", Journal of The Korean Society for Computer Game, Vol. 23, pp. 79-85, 2010.
H. J. Na, Y. E. Choe, and M. J. Chung, "Efficient 3D terrain mapping based on normal distribution transform grid", Journal of ICCAS, Vol. 76, No. 16, pp. 656-660, 2014.
N. Yokoya, K. Yamamoto, and N. Funakuro, "Fractal-based analysis and interpolation of 3D natural surface shapes and their application to terrain modeling", Computer Vision, Graphics, Image Processing, Vol. 46, No. 3, pp. 284-302, 1989. 10.1016/0734-189X(89)90034-0
고정운, 경병표, 유석호, 이동열, 이완복, 이동엽, "절차적 생성 알고리즘을 이용한 3차원 게임월드 제작", 산업융합연구 , 제16권 제1호, pp. 35-40, 2018.
https://ko.wikipedia.org/wiki/L-system
https://studyfield.tistory.com/487
https://oskarstalberg.com/game/wave/wave.html
https://velog.io/@1217pgy/
http://www.tcpschool.com/deep2018/deep2018_deeplearning_intro
https://ebbnflow.tistory.com/119
https://www.jobindexworld.com/contents/view/4379

Information

Publisher :The Society of Convergence Knowledge
Publisher(Ko) :융복합지식학회
Journal Title :The Society of Convergence Knowledge Transactions
Journal Title(Ko) :융복합지식학회논문지
Volume : 11
No :3
Pages :113-122
DOI :https://doi.org/10.22716/sckt.2023.11.3.030

[1] Newzoo, "2023 Global Games Market Report", 2023.

[2] KOCCA, "2022 Korean Game White Paper", 2023.

[3] https://newsinside.tistory.com/38 , 2022.

[4] KOCCA, "글로벌 게임산업 트렌드(2023년 1+2월호)", 2023.

[5] Y. R. Sim, H. Y. Kang, and J. H. Han, "Multi-resolution Perlin Noise Decomposition and Procedural Texture Synthesis by Example", In Conference on HCI Korea, Vol. 2016, No. 1, pp. 382-388, 2016.

[6] J. W. Ko, B. P. Kyung, and D. Y. Lee, "Procedural Map Generator Algorithm for Variableness Game World", In Conference on Korea Game Society, pp. 159-162, 2017.

[7] C. J. Hart, "Perlin noise pixel shaders", Proceedings of the ACM SIGGRAPH/EUROGRAPHICS workshop on Graphics hardware. pp. 87-94, 2001. 10.1145/383507.383531 11172120 PMC4096700

[8] K. M. Jeong, "Implementation of a 3D Terrain Map Editor Based on Height Map", Journal of The Korean Society for Computer Game, Vol. 23, pp. 79-85, 2010.

[9] H. J. Na, Y. E. Choe, and M. J. Chung, "Efficient 3D terrain mapping based on normal distribution transform grid", Journal of ICCAS, Vol. 76, No. 16, pp. 656-660, 2014.

[10] N. Yokoya, K. Yamamoto, and N. Funakuro, "Fractal-based analysis and interpolation of 3D natural surface shapes and their application to terrain modeling", Computer Vision, Graphics, Image Processing, Vol. 46, No. 3, pp. 284-302, 1989. 10.1016/0734-189X(89)90034-0

[11] 고정운, 경병표, 유석호, 이동열, 이완복, 이동엽, "절차적 생성 알고리즘을 이용한 3차원 게임월드 제작", 산업융합연구 , 제16권 제1호, pp. 35-40, 2018.

[12] https://ko.wikipedia.org/wiki/L-system

[13] https://studyfield.tistory.com/487

[14] https://oskarstalberg.com/game/wave/wave.html

[15] https://velog.io/@1217pgy/

[16] http://www.tcpschool.com/deep2018/deep2018_deeplearning_intro

[17] https://ebbnflow.tistory.com/119

[18] https://www.jobindexworld.com/contents/view/4379

The Society of Convergence Knowledge Transactions ISSN:2287-8920(Print) 융복합지식학회논문지

All Issue