AI-based Application for Drawing Traditional Thai Line Art

Abstract

Traditional Thai painting is a form of traditional art that holds significant cultural value in Thailand, particularly Thai patterns that are commonly found in various artistic works such as mural paintings, carvings, and religious ornaments. However, creating Thai patterns requires advanced skills and extensive experience, making it time-consuming for those interested in learning or practicing. This research introduces the application of artificial intelligence to develop an application that can automatically generate Thai pattern images using the Stable Diffusion Model combined with the LoRA (Low-Rank Adaptation) technique. This system is designed to promote learning and preserve traditional Thai art in the digital era while reducing skill and time barriers in the learning process.

The experimental results demonstrate that the system effectively generates Thai pattern images with high detail and strong resemblance to traditional Thai art. Additionally, the system enables users to learn and create Thai patterns more easily.

The research findings indicate that 1) the design and development of the Stable Diffusion Model integrated with the LoRA technique achieved a Fréchet Inception Distance (FID) score of 253.08—lower than the baseline Stable Diffusion v1.5 (FID = 302.09)—indicating improved image realism, and a CLIP similarity score of 0.57, reflecting stronger semantic alignment between the generated images and text prompts. 2) A comparison of user satisfaction before and after using the application was conducted with a sample group of 10 participants using a paired t-test. The results showed a statistically significant difference at the 0.05 level, with post-application satisfaction (M = 4.74, SD = 0.14) being higher than pre-application satisfaction (M = 3.25, SD = 0.45) in all aspects. The greatest improvement was observed in efficiency (t = -10.25), followed by design (t = -8.98) and usability (t = -8.52), respectively. These statistical findings demonstrate not only the technical efficiency of the system but also its effectiveness in supporting cultural preservation and promoting the learning of traditional Thai art in the digital era.

The findings highlight that the developed application effectively supports and promotes the preservation of Thai traditional art in the digital age.