Cat Emotional Sound Recognition System

This study proposes a cat emotion sound recognition method based on Gaussian Mixture Model (GMM) and Mel-frequency cepstral Coefficient (MFCC), aiming to solve the problems of insufficient data samples and single feature extraction in the classification of animal emotion sounds. By collecting 50 cat sound samples covering five types of emotions including “anger”, “happiness”, “hunting”, “unease” and “calling for kittens”, combined with MFCC feature extraction and GMM classification model, high-precision recognition of the basic emotions of cats was achieved. The experimental results show that except for the “baby-calling” emotion, which has a relatively low accuracy rate (33%) due to its highly similar spectral characteristics to “unease”, the recognition accuracy rates of the other emotions all reach 100%. This study is the first to apply GMM to the field of cat emotional sound classification, verifying the effectiveness of traditional machine learning methods in small sample scenarios and providing technical references for the development of intelligent pet devices and animal medical diagnosis.