BROWSE
Detail
Conference Paper
Prompt-Based Segmentation and Inpainting : A New Approach to Disaster Image Creation
Cited 0 time in 
Share 
Share
- Authors
- Minji Choi, Ru-Bin Won, Ji Hoon Choi, Byungjun Bae
- Issue Date
- 2023-10
- Citation
- International Conference on Information and Communication Technology Convergence (ICTC) 2023, pp.1-4
- Publisher
- IEEE
- Language
- English
- Type
- Conference Paper
- Abstract
- This paper proposes a novel approach to disaster image generation using prompt-based segmentation techniques. By segmenting terrains based on the provided prompt and inputting disaster-related prompts into the segmented area, we explore a method to generate images that reflect disaster scenarios. Moreover, we propose a method to adjust the inpainting mask area according to the severity of the disaster, providing a visual representation that varies with the situation. We acknowledge limitations in areas such as the inpainting mask region and the overall disaster image generation, and suggest directions for further research to overcome these challenges. We emphasize that the methodology of our study contributes significantly to disaster management and information dissemination.
- KSP Keywords
- Disaster Scenarios, Disaster management, Information Dissemination, New approach, Novel approach, Segmentation techniques, Visual Representation, image generation