Misophonia Trigger Sound Detection on Synthetic Soundscapes Using a Hybrid Model with a Frozen Pre-Trained CNN and a Time-Series Module

Researchers have developed a new AI model to detect sounds that trigger distress for people with misophonia, a condition causing strong negative reactions to noises like chewing. Using synthetic audio data, they tested a hybrid system combining a pre-trained feature extractor with a time-series module. The best-performing model used a Bidirectional GRU, achieving high accuracy. A lightweight model also showed promise for easy personalization with just a few user-provided sound clips.