Chord-Color Mapping for Audio-Visual Generation: Integrating Deep Learning and Emotion Quantification

This study explores chord-color mapping as a means of enhancing AI-driven audio-visual art. Addressing the longstanding lack of systematic analysis of the emotional connection between chords and colors, we propose a model that integrates deep learning and emotion quantification. The model establishes a "color-emotion-chord" pathway by leveraging image color features and a chord dataset with emotional labels, trained using an LSTM architecture for emotion-driven audio-visual generation. Evaluations demonstrate the model’s ability to substantially enhance emotional expression and aesthetic appeal while preserving essential image information. This research contributes new methodologies for cross-sensory art, music therapy, interactive design, and art education.