TY - GEN
T1 - Identifying individual facial expressions by deconstructing a neural network
AU - Arbabzadah, Farhad
AU - Montavon, Grégoire
AU - Müller, Klaus Robert
AU - Samek, Wojciech
N1 - Funding Information:
This work was supported by the German Ministry for Education and Research as Berlin Big Data Center BBDC (01IS14013A), the Deutsche Forschungsgesellschaft (MU 987/19-1) and the Brain Korea 21 Plus Program through the National Research Foundation of Korea funded by the Ministry of Education. Correspondence to KRM and WS.
Publisher Copyright:
© Springer International Publishing AG 2016.
PY - 2016
Y1 - 2016
N2 - This paper focuses on the problem of explaining predictions of psychological attributes such as attractiveness, happiness, confidence and intelligence from face photographs using deep neural networks. Since psychological attribute datasets typically suffer from small sample sizes, we apply transfer learning with two base models to avoid overfitting. These models were trained on an age and gender prediction task, respectively. Using a novel explanation method we extract heatmaps that highlight the parts of the image most responsible for the prediction. We further observe that the explanation method provides important insights into the nature of features of the base model, which allow one to assess the aptitude of the base model for a given transfer learning task. Finally, we observe that the multiclass model is more feature rich than its binary counterpart. The experimental evaluation is performed on the 2222 images from the 10k US faces dataset containing psychological attribute labels as well as on a subset of KDEF images.
AB - This paper focuses on the problem of explaining predictions of psychological attributes such as attractiveness, happiness, confidence and intelligence from face photographs using deep neural networks. Since psychological attribute datasets typically suffer from small sample sizes, we apply transfer learning with two base models to avoid overfitting. These models were trained on an age and gender prediction task, respectively. Using a novel explanation method we extract heatmaps that highlight the parts of the image most responsible for the prediction. We further observe that the explanation method provides important insights into the nature of features of the base model, which allow one to assess the aptitude of the base model for a given transfer learning task. Finally, we observe that the multiclass model is more feature rich than its binary counterpart. The experimental evaluation is performed on the 2222 images from the 10k US faces dataset containing psychological attribute labels as well as on a subset of KDEF images.
UR - http://www.scopus.com/inward/record.url?scp=84988322443&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-45886-1_28
DO - 10.1007/978-3-319-45886-1_28
M3 - Conference contribution
AN - SCOPUS:84988322443
SN - 9783319458854
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 344
EP - 354
BT - Pattern Recognition - 38th German Conference, GCPR 2016, Proceedings
A2 - Andres, Bjoern
A2 - Rosenhahn, Bodo
PB - Springer Verlag
T2 - 38th German Conference on Pattern Recognition, GCPR 2016
Y2 - 12 September 2016 through 15 September 2016
ER -