Abstract
Purpose: :
Large archives of film-based fundus photographs from completed clinical trials create a significant storage burden for the Reading centers. The image data are invaluable resource for constant knowledge extraction, but often remain unutilized due to restriction in image accessibilities. We sought to develop and evaluate an efficient system for digital archiving of film-based fundus images.
Methods: :
The design included selecting and testing of commercially available robust film digitizer, building a fast computer system and developing of software interface. Only one film scanner from all major manufacturers was identified of being capable to incorporate automatic image separation. Windows OS based computer with two Xeon processors and 1GB RAM was custom build to accommodate simultaneous use of two film scanners. The software was developed with fast programming languages. It interfaces the scanners and completely automates the acquisition, image processing, labeling and storage of digitized images. 200 film images were processed by the system at the initial evaluation.
Results: :
The system was tested for speed, reliability, accuracy, reproducibility.The mean time for processing of a single color image from 200 scans was 27 sec. and 13 sec. for black and white images. Scanning time was found to be a function of the computer recourses, scanner parameters and software optimization. The system yielded 100 % reliability at the time of testing. The digital output maintained 100 % accuracy and reproducibility for image resolution, file compression, image and file labeling.
Conclusions: :
Efficient archiving system for digitizing of very large film-based data-sets was developed and successfully validated for the purposes of clinical trials.
Keywords: clinical (human) or epidemiologic studies: systems/equipment/techniques • image processing • retina