A key characteristic of research at the Pattern Recognition Lab is its embedding in multidisciplinary environments that ensure relevance and facilitate transfer to application. This paradigm is continued in a new project on enhancing OCT Angiography images supervised by Prof. Andreas Maier, which is embedded in an ongoing collaboration with engineers at the Massachussets Institute of Technology supervised by Prof. James Fujimoto and ophthalmologists at the New England Eye Center under supervision of Prof. Nadia Waheed and Prof. Jay Duker. The German research foundation supports the researchers with about 175.000 Euro over the course of two years.

Various eye diseases show first manifestations in changes of the microstructure or blood perfusion of the retina, long before visual impairment becomes apparent. For example, changes in retinal perfusion are studied as an early indicator in diabetic retinopathy, the most frequent cause of blindness in the middle-aged populations of Europe and North America.

Optical Coherence Tomography (OCT) Angiography is a recent imaging modality that enables in-vitro 3-D visualization of blood flow, therefore allowing, in contrast to prior imaging modalities, the separation of various retinal and choroidal vessel trees and their distinct analysis. Clinical application faces two major challenges: since the angiography computation is based on differences of OCT signals acquired from the same location, it is inherently noisy and the acquisition time of a full volume is further increased. This is problematic because eye motion, which is involuntarily performed to maintain vision, leads to distortions in the scan and thus degrades thereof derived disease metrics.

Project DFG MA 4898/12-1 “Joint Iterative Reconstruction and Motion Compensation for Optical Coherence Tomography Angiography” aims at precisely modeling the characteristics of OCT Angiography scans to a) achieve reliable motion correction at scan frequency and system resolution accuracy, and b) harness the potential of compressed sensing iterative reconstruction methods on improving image quality, by transferring approaches from other clinical modalities like MRI to the domain of OCT Angiography.

Further information

Stefan Ploner, M.Sc.

http://www5.cs.fau.de/~ploner