Spectrally-based fundus imaging: implications for image enhancement and diagnosis of retinal diseases

This paper demonstrates the utility of spectral imaging for improvement in visualization of pathologic lesions and normal anatomical tissue. Traditional color images produced by current film and digital cameras are limited in the information that is provided to the physician for detecting and subsequently diagnosing retinal disorders. A high-resolution spectral fundus imaging system was developed and tested on human subjects. Spectral images were found to provide new insight into the lesions usually classified by ophthalmologists in terms of color and morphology. Analysis of spectral data had strong correlation to the qualitative classification of drusen; however, there appeared to be subclasses of the same drusen. This finding may lead to a redefinition of drusen classes that could improve correlation of drusen to the risk assessment in epidemiological studies.