Open AccessValidation of the Image Registration Technique from Functional Near Infrared Spectroscopy (fNIRS) Signal and Positron Emission Tomography (PET) Image
Author(s) -
Fairuz Mohd Nasir,
Hiroshi Watabe,
Cyclotron Safety Control
Publication year - 2020
Publication title -
international journal of management and humanities
Language(s) - English
Resource type - Journals
ISSN - 2394-0913
DOI - 10.35940/ijmh.i0877.054920
Subject(s) - imaging phantom , functional near infrared spectroscopy , positron emission tomography , image registration , artificial intelligence , computer science , computer vision , medical physics , nuclear medicine , physics , medicine , image (mathematics) , prefrontal cortex , cognition , psychiatry
Functional near infrared spectroscopy (fNIRS) is an imaging system that can measure hemodynamic changes of the brain. However, the system incapability to measure beyond the brain cortex region make it usage less appealing for in-depth brain studies. To overcome this, many researchers combine fNIRS with other imaging modalities to gain better understanding of the brain activities. In this paper, we described the theory of the registering fNIRS signals and positron emission tomography (PET) image method and performed experiments to validate it. The registration method was validated using specially designed phantom for fNIRS and PET. Polaris system was used to track the position of the phantom which is based on the Polaris markers during fNIRS and PET procedures. The Polaris markers share the same coordinate, thus the fNIRS and PET were calibrated to each other through these markers. To register the fNIRS signal on the PET image, the phantom position in fNIRS coordinate is translated to PET coordinate which allow the probe and the markers being coordinated in PET. Polaris markers were used as the references marker to determine the transformation matrices. The result shows that the fNIRS channel can be viewed on the PET image of the phantom. The transformation error from Polaris to PET is less than 1.00 mm and the precision test is less than 0.1mm while the accuracy is less than 2.8 mm. This result suggests that our theory on the registration method could be used for multimodal image registration between fNIRS and other modalities.