Abstract
The Uncanny Valley (UV) is a vital part of design research because it directly affects users' emotional responses and acceptance of anthropomorphic technical products. Traditional research relies on curve fitting to measure UV effects. However, these works often overlook the impact of data quality including scale and distribution on the accuracy and stability of fitting results. This study places a strong emphasis on the mediating role of data in UV, revisiting UV using a dataset comprising 1,000 static facial images of humanoid entities, evenly spanning the entire human likeness spectrum. The results reveal a different UV shape than Mori's original curve, especially for humanoid entities with moderate to low human likeness. Additionally, this paper explores how data quality affects UV effect curve fitting results by using sampling technologies to construct subsets. We highlight the importance of data-driven design research and provide a new perspective on avoiding and alleviating UV effects.
Keywords
uncanny valley; data-driven; design research; curve-fitting
DOI
https://doi.org/10.21606/drs.2024.422
Citation
Li, X., Xiao, Y., Zheng, Y., Qiao, J., and Leung, C. (2024) Revisiting the Uncanny Valley Effect: A data-driven analysis with curve fitting perspective, in Gray, C., Ciliotta Chehade, E., Hekkert, P., Forlano, L., Ciuccarelli, P., Lloyd, P. (eds.), DRS2024: Boston, 23–28 June, Boston, USA. https://doi.org/10.21606/drs.2024.422
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
Conference Track
Research Paper
Included in
Revisiting the Uncanny Valley Effect: A data-driven analysis with curve fitting perspective
The Uncanny Valley (UV) is a vital part of design research because it directly affects users' emotional responses and acceptance of anthropomorphic technical products. Traditional research relies on curve fitting to measure UV effects. However, these works often overlook the impact of data quality including scale and distribution on the accuracy and stability of fitting results. This study places a strong emphasis on the mediating role of data in UV, revisiting UV using a dataset comprising 1,000 static facial images of humanoid entities, evenly spanning the entire human likeness spectrum. The results reveal a different UV shape than Mori's original curve, especially for humanoid entities with moderate to low human likeness. Additionally, this paper explores how data quality affects UV effect curve fitting results by using sampling technologies to construct subsets. We highlight the importance of data-driven design research and provide a new perspective on avoiding and alleviating UV effects.