Current Result Document :
| 한글제목(Korean Title) |
심층 인공신경망의 착시 영상에 대한 반응 분석 |
| 영문제목(English Title) |
Response Analysis of Deep Neural Networks to Optical Illusion Video |
| 저자(Author) |
이강훈
김기범
이민수
장병탁
Ganghun Lee
Kibeom Kim
Minsu Lee
Byoung-Tak Zhang
|
| 원문수록처(Citation) |
VOL 48 NO. 02 PP. 0804 ~ 0806 (2021. 12) |
한글내용
(Korean Abstract) |
|
영문내용
(English Abstract) |
An optical illusion is common in humans and animals, as a by-product of brain activities. If the optical illusion is an outcome of visual information processing in the brains with evolutionary bias, how would it be for artificial neural networks? In this paper, we build and train two types of deep neural network (DNN)-based vision models, then examine whether they are fooled by the prepared optical illusion videos. Experimental results show that they also felt optical illusions, but sometimes in a way different from humans. The results mainly suggest an implication that the optical illusions are induced more by the developmental process of organisms rather than the hard-wired evolutionary priors. |
| 키워드(Keyword) |
|
| 파일첨부 |
PDF 다운로드
|
