Chiao-Ying LinJYH-HORNG CHEN2018-09-102018-09-102012-101939800Xhttp://scholars.lib.ntu.edu.tw/handle/123456789/374049https://www.scopus.com/inward/record.uri?eid=2-s2.0-84878969980&doi=10.1121%2f1.4767964&partnerID=40&md5=14d87242f0b9069fea40e5b70909144fMagnetic resonance imaging (MRI) is extensively used in clinical and medical researches because MRI is non-invasive and non-radiation. However, using the MRI cause the loud acoustic noise generated by gradient switching. The MRI noise is annoying patients. Therefore, active noise cancellation (ANC) is used to solve this problem in this article. In this article, the main work is to implement real-time system using DSP. The noise generated by echo planar imaging (EPI), which is generally used in fMRI researches, is recorded from 3 Tesla MRI for testing ANC system. The EPI signal is regular and predictable, means the signal like a pattern, we using LMS algorithm to training the signal and learning predict the noise spectrum. After the calculated the noise, we output the reverse single to we home-built headset. © 2012 Acoustical Society of America.Active noise cancellation; Echo planar imaging; Gradient switching; Home-built; LMS algorithms; Magnetic Resonance Imaging (MRI); Medical research; Noise spectra; Magnetic resonance imaging; Real time systems; Acoustic noiseconference paper10.1121/1.47679642-s2.0-84878969980