FULL ARTICLE Histogram-based filtering for quantitative 3D retinal angiography Christoph Kolbitsch, Tilman Schmoll, and Rainer A. Leitgeb * Center for Biomedical Engineering and Physics, Medical University of Vienna, Waehringerstrasse 13, A-1090 Vienna, Austria Received 23 March 2009, revised 14 May 2009, accepted 15 May 2009 Published online 15 June 2009 Key words: ocular blood flow, velocity map, Doppler OCT, FDOCT, 3D data visualisation PACS: 42.66.–p, 47.63.Jd, 47.80.Jk, 87.19.U–, 87.63.lt, 87.85.Pq # 2009 by WILEY-VCH Verlag GmbH& Co. KGaA, Weinheim 1. Introduction Histograms are tools frequently used to analyse and improve image data. They can be used for exposure corrections, feature retrieval and even for noise fil- tering [1]. Here we want to present a new field of application; a way of segmenting information in Doppler Fourier Domain optical coherence (D- FDOCT) tomograms. D-FDOCT provides information about the di- rection and the velocity of ocular blood flow [2, 3]. Doppler maps of ocular blood flow can be a valu- able tool in helping to diagnose illnesses like macula degeneration or diabetic retinopathy. These medical conditions can cause the blood vessels in the retina to clog, burst or form adenoids, which eventually leads to a partial and irreversible loss of vision. Early diagnosis is therefore essential to successful # 2009 by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Journal of BIOPHOTONICS Doppler Fourier domain optical coherence tomography (D-FDOCT) can be used to visualise and measure ocu- lar blood flow in the retina of the human eye. This has the potential to diagnose illnesses such as macula degen- eration or diabetic retinopathy early on because these medical conditions cause a pathological change in the retinal vasculature and perfusion. Here we are present- ing a method to separate quantitative blood flow infor- mation from static tissue. Our approach is demonstrated on retinal D-FDOCT volume scans taken at the optic nerve head (see figure) and near the fovea with a high speed CMOS-based FDOCT system. The advantage of this method is the small post-processing effort together with the immediate availability of segmented quantita- tive Doppler flow maps. The performance of the filter- ing technique is also compared to spatial frequency fil- tering methods. Overlay of intensity (greyscale) and blood flow velocity (color-coded) data of a FDOCT volume scan around the optic nerve head. * Corresponding author: e-mail: rainer.leitgeb@meduniwien.ac.at, Tel.: +43 1 4277 607 27, Fax.: +43 1 4277 9607 J. Biophoton. 2, No. 6 – 7, 416 – 425 (2009) / DOI 10.1002/jbio.200910026