Image quality metrics for optical coherence angiography

Biomed Opt Express. 2015 Jun 12;6(7):2435-47. doi: 10.1364/BOE.6.002435. eCollection 2015 Jul 1.


We characterized image quality in optical coherence angiography (OCA) en face planes of mouse cortical capillary network in terms of signal-to-noise ratio (SNR) and Weber contrast (Wc) through a novel mask-based segmentation method. The method was used to compare two adjacent B-scan processing algorithms, (1) average absolute difference (AAD) and (2) standard deviation (SD), while varying the number of lateral cross-sections acquired (also known as the gate length, N). AAD and SD are identical at N = 2 and exhibited similar image quality for N<10. However, AAD is relatively less susceptible to bulk tissue motion artifact than SD. SNR and Wc were 15% and 35% higher for AAD from N = 25 to 100. In addition data sets were acquired with two objective lenses with different magnifications to quantify the effect of lateral resolution on fine capillary detection. The lower power objective yielded a significant mean broadening of 17% in Full Width Half Maximum (FWHM) diameter. These results may guide study and device designs for OCA capillary and blood flow quantification.

Keywords: (100.0100) Image processing; (110.4500) Optical coherence tomography; (170.1470) Blood or tissue constituent monitoring; (170.3880) Medical and biological imaging.