A COMPARISON OF NOISE CANCELLATION OF SPEECH SIGNALS USING ADAPTIVE LMS ALGORITHM AND DEEP LEARNING ALGORITHM

Electronics and Communication Engineering

In the perspective view of communication system, noise can be defined as any undesired signal that is present in the system. The amount of noise should be reduced to a significant extend for proper transmission and reception of the signals. There are various types of methods and algorithms to remove different types of noises in which the background noise cancellation algorithms are of a major concern. This project proposes a novel comparison of the two majorly used noise removal techniques namely Adaptive noise cancellation using Least Mean Square filters and noise cancellation using Deep Learning Networks (Convolutional Neural Networks).The paper also estimates which method is more suitable for noise cancellation. Stimulation results shows that noise cancellation using deep neural networks is found to retrieve the messages more efficiently than that compared to the traditional Least Mean Square filters. The comparative analysis is done on the basis of parameters like Signal to Noise Ratio, Average power and Frequency response.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.