Images generated by my 'Kansai-embedded' Stable Diffusion XL models, the models are trained on 450 photos captured during my 2024 field research in the Japanese Kansai region and 30 satellite maps of the neighborhoods I visited.
