Automatic speaker verification has many potential applications in security, surveillance and access control. In many of these applications, it is necessary to verify the speaker based on a short and noise degraded speech utterance. This thesis addresses the problem of robust speaker verification in environmental noise conditions by introducing novel and computationally efficient techniques that are suitable for realistic conditions. It also engenders the application of psychoacoustics to realize an adaptive model compensation technique. age, by 22% over the best performing white noise based multi-conditioning technique. Moreover, the computational complexity of the proposed multiconditioning technique is also significantly reduced due to the avoidance of compute-intensive posterior union model. Finally, the proposed techniques should pave the way for realizing speaker-aware human-computer interactions in mass volume products. Contents List of Figures viii List of Tables x Glossary xi SS Spectral Subtraction; A feature domain noise robustness technique TIMIT TI-MIT(Texas Instruments-Massachusetts Institute of Technology); A speech database with high quality clean speech UBM Universal Background Model; A background model for speaker verification xii