Speech recognition using data analysis and dilation of speech content from separated audio input

A system for speech recognition using dilation of speech content from a separated audio input, which comprises:

• a computer system comprising;
• a computer processor, a computer-readable storage medium, and program instructions stored on the computer-readable storage medium being executable by the processor, to cause the computer system to perform the following functions to;
• receive an audio input, and receiving predicted changes for the audio input based on an external noise, at a CNN (Convolutional Neural Network), the audio input having speech content from a speaker;
• apply, in the CNN, diarization to the audio input to predict how a dilation of speech content from the speaker changes the audio input to generate a CNN output;
• determine a resulting dilation from the CNN output, the resulting dilation of the CNN output includes separating the sounds of the audio input;
• converting the sounds to text using the CNN;
• determine a word error rate for the dilated CNN output, using the text, to determine an accuracy for speech to text outputs;
• set an adjustment parameter to change a range of the dilation based on the word error rate;
• and
• adjust the resulting dilation of the CNN output based on the adjustment parameter to reduce the word error rate.

A computer program product for speech recognition using dilation of speech content from a separated audio input, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform functions, by the computer, comprising the functions to:

• receive an audio input, and receiving predicted changes for the audio input based on an external noise, at a CNN (Convolutional Neural Network), the audio input having speech content from a speaker;
• apply, in the CNN, diarization to the audio input to predict how a dilation of speech content from the speaker changes the audio input to generate a CNN output;
• determine a resulting dilation from the CNN output, the resulting dilation of the CNN output includes separating the sounds of the audio input;
• converting the sounds to text using the CNN;
• determine a word error rate for the dilated CNN output, using the text, to determine an accuracy for speech to text outputs;
• set an adjustment parameter to change a range of the dilation based on the word error rate;
• and
• adjust the resulting dilation of the CNN output based on the adjustment parameter to reduce the word error rate.