EECS 351: Intro to Digital Signal Processing
FINAL PROJECT

 NOISE CANCELLATION

Group 16 - Qianxu Li, Tianwei Liu, Raj Patel

DigSigProc_transp300_UToronto.png

Figure: EECS 351 Course Image © EECS 351 Course Files

 

INTRODUCTION

This website is created for presenting the outcomes of  the final project of EECS 351: Intro to Digital Signal Processing class at University of Michigan by Group 16 - Qianxu Li, Tianwei Liu, and Raj Patel.

Motivation

Inspired by the importance of audio and video recording over electronics today, our task focuses on the removal of audible noise from audio signals. In audio and video recordings, noises have different features that could be extracted using different DSP tools. Our task focuses on recordings of human voice, such as audio or videos, that are mixed with certain types of noisy components, such as train noise, people talking in the background, or machine working. And we are to use DSP skills such as frequency analysis, frequency filtering, change of basis, wave decomposition, independent component analysis, mel frequency cepstral coefficient, and etc. to separate, remove noise, and reconstruct the audio signal with clearer human voice.

Task

The goal of this project is to create a system that can effectively solve cock-tail party problems by separating each speaker and remove noise accordingly. This system can be used in various applications such as auto caption, noise cancellation for zoom, real-time audio selection, and etc.

ABOUT US

Qianxu Li

  • Third-year EE major

  • I am interested in digital integrated circuits and digital signal processing (DSP). I believe that the DSP skills can help me thoroughly analyze the signals outputted from real life applications such as the integrated circuits, so that I can make changes accordingly to enhance performance.

  • I can be contacted at qianxuli@umich.edu

 

Tianwei Liu

  • Second-year CE major

  • I am currently interested in embedded systems and system software. And I am also exploring other possibilities in the massive field of EECS. Using DSP skills, I would like to design and implement hardware systems and software algorithms that could be integrated together to solve real world problems in the future.

  • I can be contacted at twliu@umich.edu

 

Raj Patel

  • Fourth-year BME major

  • I am majorly interested in biomechanics and biocomputational. I enjoyed the class and the materials we covered. I’m excited to apply DSP and other material we learned through the semester to biomedical related problems starting with the EMG signal analysis with the NeuroProsthetics Teams here at University of Michigan.

  • I can be contacted at rajkp@umich.edu