DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Kwon, Dong-soo | - |
dc.contributor.advisor | 권동수 | - |
dc.contributor.author | Shin, Minchul | - |
dc.date.accessioned | 2018-06-20T06:15:09Z | - |
dc.date.available | 2018-06-20T06:15:09Z | - |
dc.date.issued | 2017 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=675110&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/242845 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 기계공학과, 2017.2,[iii, 37 p. :] | - |
dc.description.abstract | A built-in environment in a dataset plays an important role to decide the performance of a classifier. Up until recently, many facial expression recognition algorithms have competed their performances on a benchmark dataset. However, here one question arises. Does a classifier best in a single benchmark dataset work really better in the real-world environment? To design a classifier working robustly in the real-world, we present an Environment-diversified Network(EdNet). EdNet is a committee of a diverse dataset-biased members which share the feature extraction layers, and 90 of dataset-biased members were trained on 15 blended datasets. Rather than beating the state-of-the-art accuracy on the benchmark dataset, we focused on reducing the accuracy loss of a classifier under unfamiliar environment which the classifier have not been trained on. Finally, we confirmed that EdNet can achieve outperforming cross-dataset generalization by having diversified dataset-biased members. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | emotion | - |
dc.subject | facial expression | - |
dc.subject | deep learning | - |
dc.subject | CNN | - |
dc.subject | dataset | - |
dc.subject | bias | - |
dc.subject | real-world | - |
dc.subject | environment | - |
dc.subject | committee machine | - |
dc.subject | 얼굴 표정 인식 | - |
dc.subject | 감정 인식 | - |
dc.subject | 딥 러닝 | - |
dc.subject | 데이터셋 | - |
dc.subject | 편향 | - |
dc.subject | 심층 신경망 | - |
dc.subject | 실환경 | - |
dc.subject | 커미티머신 | - |
dc.subject | 앙상블 | - |
dc.title | Committee of dataset-biased CNNs for real-world application of facial expression recognition | - |
dc.title.alternative | 환경변화에 강인한 얼굴표정 인식을 위한 심층 나선형 신경망 기반 환경 편향 커미티머신 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :기계공학과, | - |
dc.contributor.alternativeauthor | 신민철 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.