TY - GEN
T1 - FaceNet2ExpNet
T2 - 12th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2017
AU - Ding, Hui
AU - Zhou, Shaohua Kevin
AU - Chellappa, Rama
N1 - Publisher Copyright:
© 2017 IEEE.
PY - 2017/6/28
Y1 - 2017/6/28
N2 - Relatively small data sets available for expression recognition research make the training of deep networks very challenging. Although fine-tuning can partially alleviate the issue, the performance is still below acceptable levels as the deep features probably contain redundant information from the pretrained domain. In this paper, we present FaceNet2ExpNet, a novel idea to train an expression recognition network based on static images. We first propose a new distribution function to model the high-level neurons of the expression network. Based on this, a two-stage training algorithm is carefully designed. In the pre-training stage, we train the convolutional layers of the expression net, regularized by the face net; In the refining stage, we append fully-connected layers to the pre-trained convolutional layers and train the whole network jointly. Visualization results show that the model trained with our method captures improved high-level expression semantics. Evaluations on four public expression databases, CK+, Oulu- CASIA, TFD, and SFEW demonstrate that our method achieves better results than state-of-the-art.
AB - Relatively small data sets available for expression recognition research make the training of deep networks very challenging. Although fine-tuning can partially alleviate the issue, the performance is still below acceptable levels as the deep features probably contain redundant information from the pretrained domain. In this paper, we present FaceNet2ExpNet, a novel idea to train an expression recognition network based on static images. We first propose a new distribution function to model the high-level neurons of the expression network. Based on this, a two-stage training algorithm is carefully designed. In the pre-training stage, we train the convolutional layers of the expression net, regularized by the face net; In the refining stage, we append fully-connected layers to the pre-trained convolutional layers and train the whole network jointly. Visualization results show that the model trained with our method captures improved high-level expression semantics. Evaluations on four public expression databases, CK+, Oulu- CASIA, TFD, and SFEW demonstrate that our method achieves better results than state-of-the-art.
UR - http://www.scopus.com/inward/record.url?scp=85026315336&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85026315336&partnerID=8YFLogxK
U2 - 10.1109/FG.2017.23
DO - 10.1109/FG.2017.23
M3 - Conference contribution
AN - SCOPUS:85026315336
T3 - Proceedings - 12th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2017 - 1st International Workshop on Adaptive Shot Learning for Gesture Understanding and Production, ASL4GUP 2017, Biometrics in the Wild, Bwild 2017, Heterogeneous Face Recognition, HFR 2017, Joint Challenge on Dominant and Complementary Emotion Recognition Using Micro Emotion Features and Head-Pose Estimation, DCER and HPE 2017 and 3rd Facial Expression Recognition and Analysis Challenge, FERA 2017
SP - 118
EP - 126
BT - Proceedings - 12th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2017 - 1st International Workshop on Adaptive Shot Learning for Gesture Understanding and Production, ASL4GUP 2017, Biometrics in the Wild, Bwild 2017, Heterogeneous Face Recognition, HFR 2017, Joint Challenge on Dominant and Complementary Emotion Recognition Using Micro Emotion Features and Head-Pose Estimation, DCER and HPE 2017 and 3rd Facial Expression Recognition and Analysis Challenge, FERA 2017
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 30 May 2017 through 3 June 2017
ER -