2022 Fiscal Year Final Research Report
Development and Flexibility of Facial Recognition Ability: Exploring Human Traits Using Artificial Intelligence
Project/Area Number |
19K14383
|
Research Category |
Grant-in-Aid for Early-Career Scientists
|
Allocation Type | Multi-year Fund |
Review Section |
Basic Section 10020:Educational psychology-related
|
Research Institution | Kindai University (2020-2022) Osaka University (2019) |
Principal Investigator |
Ueno Masataka 近畿大学, 総合社会学部, 講師 (30737432)
|
Project Period (FY) |
2019-04-01 – 2023-03-31
|
Keywords | 霊長類 / 顔の識別 / 可塑性 / 深層学習 / 人工知能 |
Outline of Final Research Achievements |
This research project aimed to investigate the extent to which the ability to recognize others' faces is innate or acquired, and to elucidate the foundations of human sociability that stand out even among animals. In a study involving primatologists experienced in individual recognition of monkeys and other adults, it was shown that the primatologists had a higher ability to recognize novel monkey faces compared to the other adults. Furthermore, in this research project, a program for individual recognition of Japanese macaques was developed using deep learning. While deep learning enabled accurate identification of Japanese macaques, combining deep learning with sequential Bayesian filtering suggested even higher accuracy in identifying Japanese macaques.
|
Free Research Field |
比較行動学
|
Academic Significance and Societal Importance of the Research Achievements |
ヒトを含む霊長類は集団内において、他者を個体識別し、それぞれの他個体と社会関係を築く。これまでの研究では、顔の識別能力については生後1年以内の経験による影響が大きいと考えられてきた。本研究課題の研究成果から、成人した後であっても、サルを見る経験を積むことによって、サルの識別能力が向上することが示唆された。他方で、サル研究者が新奇のサル画像を見分ける場合、顔だけでなく、顔以外の身体の情報を利用していることが分かった。ここから、成人した後にも識別能力は向上するものの、乳児期に比べてその柔軟性は低い可能性が考えられた。
|