Skip to content

Latest commit

 

History

History
7 lines (5 loc) · 364 Bytes

File metadata and controls

7 lines (5 loc) · 364 Bytes

Social Robotics - Vocal Intent Recognition

Classification of audio signals based on their spectral features.
Two datasets are used and the algorithm has to classify 3 different voice patterns.
A Convolutional Neural Network is used to classify the spectrograms of the different audio samples.

Accuracy and scalability results still need to be improved.