Enhancing Emotion Recognition in Incomplete Data: A Novel Cross-Modal Alignment, Reconstruction, and Refinement Framework
Abstract: Multimodal emotion recognition systems rely heavily on the full availability of modalities, suffering significant performance declines when modal data is incomplete. To tackle this issue, we present the Cross-Modal Alignment, Reconstruction, and Refinement (CM-ARR) framework, an innovative approach that sequentially engages in cross-modal alignment, reconstruction, and refinement phases to handle missing modalities and enhance emotion recognition. This framework utilizes unsupervised distribution-based contrastive learning to align heterogeneous modal distributions, reducing discrepancies and modeling semantic uncertainty effectively. The reconstruction phase applies normalizing flow models to transform these aligned distributions and recover missing modalities. The refinement phase employs supervised point-based contrastive learning to disrupt semantic correlations and accentuate emotional traits, thereby enriching the affective content of the reconstructed representations. Extensive experiments on the IEMOCAP and MSP-IMPROV datasets confirm the superior performance of CM-ARR under conditions of both missing and complete modalities. Specifically, with all missing conditions on IEMOCAP and MSP-IMPROV, CM-ARR shows absolute improvements of 2.11% and 2.12%, 1.71% and 1.96% in WAR and UAR, respectively. Our code is available at https://anonymous.4open.science/r/CM-ARR-31C7.
Paper Type: Long
Research Area: Sentiment Analysis, Stylistic Analysis, and Argument Mining
Research Area Keywords: Sentiment Analysis,Computational Social Science and Cultural Analytics
Contribution Types: NLP engineering experiment
Languages Studied: English
Submission Number: 807
Loading