Sec2Sec Co-attention for Video-Based Apparent Affective Prediction

TL;DR

This paper introduces a novel LSTM-Transformer co-attention model for video-based affect prediction, improving accuracy and interpretability by integrating vision, audio, and spatiotemporal cues.

Contribution

It presents a new Sec2Sec Co-attention Transformer that outperforms existing methods and provides interpretability in affect prediction tasks.

Findings

Outperforms state-of-the-art on LIRIS-ACCEDE and First Impressions datasets

Provides interpretability of affective contributions over time

Effective integration of multi-modal video elements

Abstract

Video-based apparent affect detection plays a crucial role in video understanding, as it encompasses various elements such as vision, audio, audio-visual interactions, and spatiotemporal information, which are essential for accurate video predictions. However, existing approaches often focus on extracting only a subset of these elements, resulting in the limited predictive capacity of their models. To address this limitation, we propose a novel LSTM-based network augmented with a Transformer co-attention mechanism for predicting apparent affect in videos. We demonstrate that our proposed Sec2Sec Co-attention Transformer surpasses multiple state-of-the-art methods in predicting apparent affect on two widely used datasets: LIRIS-ACCEDE and First Impressions. Notably, our model offers interpretability, allowing us to examine the contributions of different time points to the overall prediction. The implementation is available at: https://github.com/nestor-sun/sec2sec.