Connectionist Spectral Mapping of Noisy Data Collected Over Telephone Channel

This work discusses experimental results obtained with a connectionist architecture, based on feed-forward and recurrent neural networks, to approach the problem of channel compensation for data collected over telephone line, through acoustic feature transformation. Recognition experiments have been carried out using speaker-independent (SI) recognition systems based on Hidden Markov Models (HMMs). The proposed approaches were hardly capable to improve baseline results when a combination of individual feature mappings was attempted