Speech Recognition On Mobile Devices In Noisy Environments

The use of speech recognition on mobile devices has been possible with the development of cloud systems and has been used for about 10 years. However, in noisy environments, the problem of speech recognition with low error rate still persists. In this study, different speech samples have been recorded using a compact microphone array in noisy environments and a data set has been created by processing them with a real-time noise cancellation algorithm. Speech recognition performance has been tested on the generated dataset using Google cloud system. As a result of the test, speech recognition performance of the cloud systems according to the noise level was observed. Results show that in order to apply speech recognition using cloud computing systems on mobile devices, the noise level has to be low or real-time noise cancellation algorithms are needed.