Emotion - Centric Palette Recommender : Synergizing Cognitive Behavioral Therapy With ML and Advanced Databases

Abstract

The rapid growth of digital technology has created numerous tools designed to improve productivity, entertainment, and social communication. However, emotional well-being remains an underexplored domain in human–computer interaction. This paper introduces an Emotion-Centric Palette Recommender System that combines Cognitive Behavioral Therapy (CBT), Machine Learning (ML), and advanced database technologies to deliver a personalized and therapeutic digital experience. The proposed system employs multimodal emotion detection by processing facial expressions, speech tones, and EEG brainwave signals to assess a user’s emotional state. A custom-built EEG sensor connected through a Bluetooth-enabled microcontroller transmits neural activity to a cloud-integrated web platform. The detected emotion triggers a color palette recommendation and a CBT-based intervention such as mindfulness prompts, journaling, or breathing exercises. A hybrid ML model integrates Convolutional Neural Networks (CNNs) for facial emotion analysis and Recurrent Neural Networks (RNNs) for temporal EEG and voice signal interpretation. An intelligent database design using Firebase and SQLite enables adaptive learning from user feedback, emotional history, and palette preferences. Experimental results reveal an overall accuracy of 94.6% for multimodal emotion detection and significant user satisfaction improvement. The system represents a novel intersection between artificial intelligence, psychology, and embedded systems, bridging emotional computing and behavioral therapy for mental wellness.