Deep learning Multimodal feature enhanced with cross-modal attention for emotion recognition
Year: 2025
Type: article
Abstract: Human emotion recognition is becoming very important in human-computer interaction applications and stress management applications. In these applications, multimodal features extracted from various modalities like visual, speech, text, and biomedical sensor readings etc. to classify various emotions. For effective emotion recognition there has been extensive research on various modalities and their features which are to be considered for classification of emotions. This work explores multiple modalities and proposes a novel deep learning based multimodal features which are enhanced by using cross modality attention for recognition of basic human emotions The proposed solution provides higher discriminative ability in three cases of emotions like basic emotion, activation of emotion (positive, negative) and arousal of emotion (high, low).
Source: Journal of Neonatal Surgery
Institution: Savitribai Phule Pune University
Language: English
Topic: Emotion and Mood Recognition
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7219336202
