Analyzing Arabic digit recognizer errors using spectrograms

In a previous paper [Yousef A. Alotaibi, 2003], Arabic digits were investigated from the speech recognition problem point of view. An artificial neural network based speech recognition system was designed and tested for automatic Arabic digit recognition. The system was an isolated whole word speech recognizer and it was implemented as both a multi-speaker and speaker-independent modes. As continuation to that research, this paper investigated Arabic digits by using spectrogram information to cross check and investigate digit recognition system errors and to try to locate the causes of miss-recognized digits. Comparisons of some possible pairs of digits were conducted and comments were stated with links to digit recognition system output.