The main objective of this work is to provide a low cost alternative for the literate deaf people suffering from Sensori Neural Loss and Central loss, or what we call as Permanent Deafness. Though Cochlear Implant is a solution for this problem, it requires the patient to spend Lakhs of rupees for the surgery. In a country like India, not all can afford such a huge amount of money. So, a low cost alternative would definitely help the middle class people. The proposed project describes Isolated Word Recognition of Kannada Digits. The system reads the spoken speech signal. Wavelet transform of the speech signal is taken and MFCC (Mel Frequency Cepstral Coefficients) are calculated followed by Vector Quantization. Euclidean Distance measure is used to correlate the test speech signal with pre recorded speech signals from the speech database. The nearest match is identified and its respective text equivalent is displayed. The project is carried out for Kannada digits, which can be extended to words later. The programming is done using Matlab