Automatic fundus image quality assessment on a continuous scale.

Fundus photography is commonly used for screening, diagnosis, and monitoring of various diseases affecting the eye. In addition, it has shown promise in the diagnosis of brain diseases and evaluation of cardiovascular risk factors. Good image quality is important if diagnosis is to be accurate and timely. Here, we propose a method that automatically grades image quality on a continuous scale which is more flexible than binary quality classification. The method utilizes random forest regression models trained on image features discovered automatically by combining basic image filters using simulated annealing as well as features extracted with the discrete Fourier transform. The method was developed and tested on images from two different fundus camera models. The quality of those images was rated on a continuous scale from 0.0 to 1.0 by five experts. In addition, the method was tested on DRIMDB, a publicly available dataset with binary quality ratings. On the DRIMDB dataset the method achieves an accuracy of 0.981, sensitivity of 0.993 and specificity of 0.958 which is consistent with the state of the art. When evaluating image quality on a continuous scale the method outperforms human raters.

Retinal oximetry with a scanning laser ophthalmoscope.

PURPOSE: The purpose of the study was to assess if a scanning laser ophthalmoscope (SLO), Optomap 200Tx, could be used for measurements of hemoglobin oxygen saturation in retinal blood vessels. METHODS: Optomap 200Tx uses two lasers for image acquisition, 532 and 633 nm. Retinal images of healthy individuals and patients with retinal vein occlusion were analyzed with modified Oxymap Analyzer software, which tracks retinal vessels and calculates relative hemoglobin oxygen saturation. RESULTS: Oxygen saturation in healthy individuals was measured as 92% ± 13% for arterioles and 57% ± 12% for venules (mean ± SD, n = 11, P = 0.0001). Standard deviation for repeated measurements of the same eye was 3.5% for arterioles and 4.4% for venules. In patients with confirmed venular hypoxia, central retinal vein occlusion (CRVO) or hemivein occlusion, the average venular oxygen saturation was measured as 23% ± 3% in the affected eyes and 59% ± 3% in the fellow eyes (n = 4, P = 0.0009). CONCLUSIONS: Technically, it is possible to derive information on retinal oxygen saturation from an SLO with a 2-wavelength oximetry algorithm. The system produced both sensitive and repeatable results. The remaining challenges include decreasing variability between vessels of the same eye and variability between individuals. Given the advantages that SLO imaging has over conventional fundus camera optics in retinal oximetry, further development of SLO oximetry may provide the optimal approach to retinal oximetry.

Spectrophotometric retinal oximetry in pigs.

PURPOSE: To assess the validity of spectrophotometric retinal oximetry by comparison to blood gas analysis and intravitreal measurements of partial pressure of oxygen (pO(2)). METHODS: Female domestic pigs were used for all experiments (n = 8). Oxygen fraction in inspired air was changed using a mixture of room air, pure oxygen, and pure nitrogen, ranging from 5% to 100% oxygen. Femoral arterial blood gas analysis and retinal oximetry were performed at each level of inspiratory oxygen fraction. Retinal oximetry was performed using a commercial instrument, the Oxymap Retinal Oximeter T1. The device simultaneously acquires images at two wavelengths (570 nm and 600 nm), and specialized software automatically detects retinal blood vessels. In three pigs, invasive pO(2) measurements were performed after the initial noninvasive measurements. RESULTS: Comparison of femoral arterial oxygen saturation and the optical density ratio over retinal arteries revealed an approximately linear relationship (R(2) = 0.74, P = 3.4 × 10(-9)). In order to test the validity of applying the arterial calibration to veins, we compared noninvasive oximetry measurements to invasive pO2 measurements in three pigs. This relationship was approximately linear (R(2) = 0.45, P = 0.04). CONCLUSIONS: Noninvasive spectrophotometric oximetry is sensitive to changes in oxygen saturation in pigs and correlated with intravitreal pO(2) measurements and with femoral artery pO(2). Pigs present a higher intraindividual variability in retinal oxygen saturation and a lower overall saturation than do humans. The difference between porcine and human eyes makes direct comparisons of measurements difficult.