Temporal Convolution Networks with Positional Encoding for Evoked Expression Estimation

TL;DR

This paper introduces a novel approach using temporal convolution networks with positional encoding to predict evoked facial expressions from videos, achieving state-of-the-art results in the EEV challenge.

Contribution

It proposes replacing RNNs with temporal convolution networks and employs positional encoding to handle missing annotations, improving evoked expression estimation.

Findings

Achieved top performance with a Pearson correlation of 0.05477.

Outperformed previous methods in the EEV 2021 challenge.

Demonstrated the effectiveness of temporal convolution networks with positional encoding.

Abstract

This paper presents an approach for Evoked Expressions from Videos (EEV) challenge, which aims to predict evoked facial expressions from video. We take advantage of pre-trained models on large-scale datasets in computer vision and audio signals to extract the deep representation of timestamps in the video. A temporal convolution network, rather than an RNN like architecture, is used to explore temporal relationships due to its advantage in memory consumption and parallelism. Furthermore, to address the missing annotations of some timestamps, positional encoding is employed to ensure continuity of input data when discarding these timestamps during training. We achieved state-of-the-art results on the EEV challenge with a Pearson correlation coefficient of 0.05477, the first ranked performance in the EEV 2021 challenge.