SRM2 Sound Recognizer Mobile 2

The SRM Framework 2 was developed for the J2ME platform supporting sound recognition dependent and speaker-independent, through the recognition techniques DTW (Dynamic Time Warping) and HMM (Hidden Markov Models), in addition to providing support to the import and export of data any stage of recognition enabling the use of external resources as well as the tool running steps on remote terminals.

It has the following characteristics:

- DTW
- HMM (Discrete Models)
- Suports Sound in PCM
- Vector quantization using the k-means algorithm
- Uses the algorithm Forward
- Uses the algorithm Backward
- Uses the Viterbi algorithm
- Uses the Baum-Welch algorithm for training
- Has features import and export data

Further information is on Gaita 2012.