The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
A microexpressão facial consiste em reações faciais momentâneas e sutis, e ainda é um desafio reconhecer automaticamente a microexpressão facial com alta precisão em aplicações práticas. Extrair características espaçotemporais de sequências de imagens faciais é essencial para o reconhecimento de microexpressões faciais. Neste artigo, empregamos Redes Neurais Convolucionais 3D (3D-CNNs) para extração de características de autoaprendizagem para representar efetivamente a microexpressão facial, uma vez que as 3D-CNNs poderiam muito bem extrair as características espaço-temporais de sequências de imagens faciais. Além disso, a aprendizagem por transferência foi utilizada para lidar com o problema de amostras insuficientes no banco de dados de microexpressões faciais. Nós pré-treinamos principalmente os 3D-CNNs no banco de dados de expressões faciais normais Oulu-CASIA por aprendizado supervisionado, então o modelo pré-treinado foi efetivamente transferido para o domínio alvo, que era a tarefa de reconhecimento de microexpressão facial. O método proposto foi avaliado em dois conjuntos de dados de microexpressão facial disponíveis, ou seja, CASME II e SMIC-HS. Obtivemos a precisão geral de 97.6% no CASME II e 97.4% no SMIC, que foram 3.4% e 1.6% maiores que o modelo 3D-CNNs sem aprendizagem por transferência, respectivamente. E os resultados experimentais demonstraram que nosso método alcançou desempenho superior em comparação aos métodos mais modernos.
Ruicong ZHI
University of Science and Technology Beijing,Beijing Key Laboratory of Knowledge Engineering for Materials Science
Hairui XU
University of Science and Technology Beijing,Beijing Key Laboratory of Knowledge Engineering for Materials Science
Ming WAN
University of Science and Technology Beijing,Beijing Key Laboratory of Knowledge Engineering for Materials Science
Tingting LI
University of Science and Technology Beijing,Beijing Key Laboratory of Knowledge Engineering for Materials Science
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Ruicong ZHI, Hairui XU, Ming WAN, Tingting LI, "Combining 3D Convolutional Neural Networks with Transfer Learning by Supervised Pre-Training for Facial Micro-Expression Recognition" in IEICE TRANSACTIONS on Information,
vol. E102-D, no. 5, pp. 1054-1064, May 2019, doi: 10.1587/transinf.2018EDP7153.
Abstract: Facial micro-expression is momentary and subtle facial reactions, and it is still challenging to automatically recognize facial micro-expression with high accuracy in practical applications. Extracting spatiotemporal features from facial image sequences is essential for facial micro-expression recognition. In this paper, we employed 3D Convolutional Neural Networks (3D-CNNs) for self-learning feature extraction to represent facial micro-expression effectively, since the 3D-CNNs could well extract the spatiotemporal features from facial image sequences. Moreover, transfer learning was utilized to deal with the problem of insufficient samples in the facial micro-expression database. We primarily pre-trained the 3D-CNNs on normal facial expression database Oulu-CASIA by supervised learning, then the pre-trained model was effectively transferred to the target domain, which was the facial micro-expression recognition task. The proposed method was evaluated on two available facial micro-expression datasets, i.e. CASME II and SMIC-HS. We obtained the overall accuracy of 97.6% on CASME II, and 97.4% on SMIC, which were 3.4% and 1.6% higher than the 3D-CNNs model without transfer learning, respectively. And the experimental results demonstrated that our method achieved superior performance compared to state-of-the-art methods.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDP7153/_p
Copiar
@ARTICLE{e102-d_5_1054,
author={Ruicong ZHI, Hairui XU, Ming WAN, Tingting LI, },
journal={IEICE TRANSACTIONS on Information},
title={Combining 3D Convolutional Neural Networks with Transfer Learning by Supervised Pre-Training for Facial Micro-Expression Recognition},
year={2019},
volume={E102-D},
number={5},
pages={1054-1064},
abstract={Facial micro-expression is momentary and subtle facial reactions, and it is still challenging to automatically recognize facial micro-expression with high accuracy in practical applications. Extracting spatiotemporal features from facial image sequences is essential for facial micro-expression recognition. In this paper, we employed 3D Convolutional Neural Networks (3D-CNNs) for self-learning feature extraction to represent facial micro-expression effectively, since the 3D-CNNs could well extract the spatiotemporal features from facial image sequences. Moreover, transfer learning was utilized to deal with the problem of insufficient samples in the facial micro-expression database. We primarily pre-trained the 3D-CNNs on normal facial expression database Oulu-CASIA by supervised learning, then the pre-trained model was effectively transferred to the target domain, which was the facial micro-expression recognition task. The proposed method was evaluated on two available facial micro-expression datasets, i.e. CASME II and SMIC-HS. We obtained the overall accuracy of 97.6% on CASME II, and 97.4% on SMIC, which were 3.4% and 1.6% higher than the 3D-CNNs model without transfer learning, respectively. And the experimental results demonstrated that our method achieved superior performance compared to state-of-the-art methods.},
keywords={},
doi={10.1587/transinf.2018EDP7153},
ISSN={1745-1361},
month={May},}
Copiar
TY - JOUR
TI - Combining 3D Convolutional Neural Networks with Transfer Learning by Supervised Pre-Training for Facial Micro-Expression Recognition
T2 - IEICE TRANSACTIONS on Information
SP - 1054
EP - 1064
AU - Ruicong ZHI
AU - Hairui XU
AU - Ming WAN
AU - Tingting LI
PY - 2019
DO - 10.1587/transinf.2018EDP7153
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2019
AB - Facial micro-expression is momentary and subtle facial reactions, and it is still challenging to automatically recognize facial micro-expression with high accuracy in practical applications. Extracting spatiotemporal features from facial image sequences is essential for facial micro-expression recognition. In this paper, we employed 3D Convolutional Neural Networks (3D-CNNs) for self-learning feature extraction to represent facial micro-expression effectively, since the 3D-CNNs could well extract the spatiotemporal features from facial image sequences. Moreover, transfer learning was utilized to deal with the problem of insufficient samples in the facial micro-expression database. We primarily pre-trained the 3D-CNNs on normal facial expression database Oulu-CASIA by supervised learning, then the pre-trained model was effectively transferred to the target domain, which was the facial micro-expression recognition task. The proposed method was evaluated on two available facial micro-expression datasets, i.e. CASME II and SMIC-HS. We obtained the overall accuracy of 97.6% on CASME II, and 97.4% on SMIC, which were 3.4% and 1.6% higher than the 3D-CNNs model without transfer learning, respectively. And the experimental results demonstrated that our method achieved superior performance compared to state-of-the-art methods.
ER -