To detect early diabetic retinopathy, we employed three-dimensional analysis of HSV colour spaces.Methods
Six patients with diabetic retinopathy and small haemorrhages, hard exudates and photocoagulation marks were evaluated using fundus photography, which revealed dust artefacts in the fundi of some patients. We constructed an experimental device similar to the optical system of a fundus camera and made the artificial eyes of the fundus painted four kinds of colours: khaki, sunset, rose and sunflower. The dust artefacts were photographed under each artificial eye using the experimental device. We analysed all images using Scilab 5.4.0 and SIVP 0.5.3 software, which changed the RGB colour space into an HSV colour space. The software interpreted each value of HSV colour space as a three-dimensional graph, which was modified using a Gaussian filter.Results
We calculated the difference between manifestation and perimanifestation areas and dust artefact and periartefact areas using average HSV values. V values for the manifestations were haemorrhage, 0.06±0.03; hard exudate, −0.12±0.06 and photocoagulation marks, 0.07±0.02. V values for the dust artefacts, visualized under the human eye and the artificial eyes, were human eye, 0.19±0.03; khaki, 0.41±0.02; sunset, 0.43±0.04; rose, 0.47±0.11 and sunflower, 0.59±0.07. Ratios of the dust artefact to manifestation areas in the human eye were haemorrhage, 3.17; hard exudate, −1.58 and photocoagulation, 2.71.Conclusion
The HSV colour space displayed good capability to distinguish small retinal haemorrhages, hard exudates and photocoagulation marks from dust artefacts.