An Exploratory Study of Kansei-Guided Reference Matching for Automotive Side-View Sketch Rendering
Year: 2026
Type: article
Abstract: To provide a systematic basis for selecting style references for automotive side-view sketches, this study presents a concept rendering workflow guided by Kansei attribute matching. The workflow ranks candidate references using ratings on six bipolar Kansei attributes and generates renderings by interpolating StyleGAN latent codes. For all six sketches, Pearson correlation and Euclidean distance identified the same top-ranked reference. Compared with renderings based on references chosen by the sketch authors, renderings based on Kansei-matched references received, on average, more fixations and longer total fixation durations in both sketch groups. Changes in author satisfaction varied across cases. ControlNet and IP-Adapter provided a supplementary comparison using the same inputs. The study offers a reference selection method whose rankings can be checked, and it demonstrates how the selected references can inform automotive concept rendering.
Source: Information
DOI: 10.3390/info17111172
Institution: Guiyang University
Language: English
Topic: Color perception and design
Subfield: Social Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $1,930
OpenAlex ID: W7236346517
