Convolutional neural networks pretrained on large face recognition datasets for emotion classification from video

Abstract

In this paper we describe a solution to our entry for the emotion recognition challenge EmotiW 2017. We propose an ensemble of several models, which capture spatial and audio features from videos. Spatial features are captured by convolutional neural networks, pretrained on large face recognition datasets. We show that usage of strong industry-level face recognition networks increases the accuracy of emotion recognition. Using our ensemble we improve on the previous best result on the test set by about 1 %, achieving a 60.03 % classification accuracy without any use of visual temporal information.

4 pages

References in corpus (3)

• YouTube-8M: A Large-Scale Video Classification Benchmark
• Challenges in Representation Learning: A report on three machine learning contests
• EmotioNet Challenge: Recognition of facial expressions of emotion in the wild

Cited by in corpus (6)

• Deep Facial Expression Recognition: A Survey
• Deep Affect Prediction in-the-wild: Aff-Wild Database and Challenge, Deep Architectures, and Beyond
• Noisy Student Training using Body Language Dataset Improves Facial Expression Recognition
• Frame attention networks for facial expression recognition in videos
• Feature-level and Model-level Audiovisual Fusion for Emotion Recognition in the Wild
• The Many Moods of Emotion