SPEECH RECOGNITION

SUMMARY

To build a speech recognition system, it is necessary to have skills in many disciplines beyond software development. Currently, there are several approaches; usually, larger companies like Google or Apple apply Neural Networks. Since we are passionate about sound, this month we experimented with a lighter approach that is applied in more restricted domains. It consists of feature extraction and DTW analysis. We uploaded a repository to GitHub with the work done; the system extracts 12 MFCC coefficients by default for the bandwidth and uses K-NN algorithms to compare the dataset.

You can download the project at the following link: https://github.com/nahuelproietto/SpeechProcessingSwift