Study on Ghibli Color Sensitivity Reproduced in AI Image Generation : An Analysis of Visual Similarities with Hayao Miyazaki Animation Posters

Since the 21st century, image generation technologies have evolved beyond imitating human intuition to structuring visual emotions through artificial intelligence (AI). This study examines how the color sensibility embedded in Studio Ghibli’s animation posters is reproduced by generative AI. We conducted a quantitative color analysis of 20 representative posters directed by Hayao Miyazaki using the HSL (Hue, Saturation, Lightness) model, and compared the results with 16 AI-generated images produced under the same conditions. The analysis revealed that Ghibli posters consistently adopt a low-saturation color strategy centered on turquoise and ocher tones to evoke emotional immersion, while AI images tended to replicate similar chromatic structures. However, limitations emerged in the narrative flow and contextual depth of emotions, as the nuanced interaction between color and emotional progression was less sophisticated than that achieved by human designers’ intuition. User evaluations also indicated that while participants perceived strong visual similarity between the original and AI-generated images, they reported lower satisfaction regarding emotional diversity. Overall, the study demonstrates that AI-generated “Ghibli-style” images go beyond mere visual imitation by forming an independent emotional structure, yet remain constrained in capturing narrative depth. These findings provide a theoretical basis for understanding both the potential and limitations of algorithmic reproduction of color sensibility, and suggest practical implications for applying generative AI technologies in emotional design.