[3B] (02-3) Education system for AI, data science and Society 5.0

2:00 PM - 2:15 PM JST(5:00 AM - 5:15 AM UTC)

[3B13] Autoencoder-Based Feature Extraction of Architectural CG Works by Students

○Kyohei NISHIOKA¹, Gakuto SAKURAI², Akinaru IINO³ (1. Niigata Institute of Technology, 2. NAO Taniyama&associates, 3. Niigata Institute of Technology)

Keywords:

Neural Network,Autoencoder,Clustering,Architectural Education,Computer Graphics

This study proposes a method to objectively analyze the latent design orientations of architecture students from their CG works. It demonstrates that students' multifaceted orientations can be visualized by integrating the classification results of an FCNN, which captures the overall atmosphere, and a CNN, which captures local textures.