Using a groundbreaking technique developed in astronomy, researchers at the University of Edinburgh (UK; www.research-innovation.ed.ac.uk) can register complete 3-D image volumes to subpixel accuracy in real time. For the first time, full affine correction can be calculated and applied as an imaging sequence is acquired, vastly reducing both bulk subject motion and geometric artifacts from the images.
The core of the technology is a fast algorithm that can speed up image processing by factors of 100 to 1,000,000. This enables processing of the extensive data from detailed 3-D images in real time. It is applicable to structural MRI, functional MRI (fMRI), and other forms of medical imaging.
The method can be used to assess image quality as the scan progresses, thereby allowing radiographers to stop and rerun the scan if it is not of suitable quality. Although the registration is currently implemented on MRI data, we are confident that the method could be extended to other imaging modalities. This technique is also applicable to accelerated drug discovery that is increasingly turning to imaging to speed up assessment of drug efficacy. Likewise there are applications in security where speed of image processing is vital.
Key benefits
* Allows real-time 3-D registration, both rigid body and full affine
* One click graphical interface operation possible
* Enables the quality of imaging data to be assessed as the sequence is acquired, thereby allowing potential rescanning while the subject is available
* Saves time, improves throughput and utilization
* Runs on commodity computing components
* Works with DICOM format images direct from an image server; no alterations to scanner required
* ISO9000:2001 compatible quality program in place to ease approval process
Applications
* Medical Imaging
* Fast Data fitting & Analysis
* Diagnostic Aids
* Quality control
* Drug discovery and efficacy monitoring
* Security
Intellectual property status
This technology enjoys protection under one granted and two pending patents and is funded by an award from the Scottish Enterprise Proof of Concept Programme.
