Abstract
Purpose: :
To develop a system for digitizing 35mm fundus slides to: 1) store images efficiently, 2) improve the preservation of clinical study data and, 3) prepare data for further image processing studies.
Methods: :
An ADCIS developed innovative system using a 10 mega-pixel digital SLR camera (approx. 2800dpi), a copy-stand, and a light box to back-light the slides. The main requirements of the system were: 1. Capture image of the slide frame with patient identifiers and associated fundus image; 2. Provide detailed index of all images for quick image filtering and retrieval; 3. Provide the user with full control of camera settings to ensure the digitized image maintained the fidelity of the slide’s image; and 4. Provide superior system performance than conventional slide scanners (typically 5 minutes/slide at ~2000dpi).The system software developed by ADCIS and Allergan Inc. meets all these requirements. It is comprised of two main components: a) Slide Snapper, to capture slide image and patient data; and b) Slide Display, to query the database and display images.Slide Snapper includes a camera driver providing the user with full control of the camera using a personal computer, and the capability to snap two images of a slide: one low resolution JPEG image of the frame, and a full resolution, RAW mode image of the film portion of the slide, later converted to a TIFF image. Slide-Snapper records images and user-entered patient data into a MySQL database for use by Slide-Display.Slide Display enables the user to query the index database with advanced selection and filtering tools to display images and data. User can adjust image brightness, contrast, zoom, and view stereo pairs (using 3D glasses). Database indexes can be exported to allow external data sources reconciliation. Index database can be edited to correct user data entry errors.
Results: :
Using two of these slide digitization systems, two trained users digitized 78,000 slides during an 81-day period. The resulting database is now online and accessible on the Allergan campus using the Slide-Display software.
Conclusions: :
35 mm fundus slides were scanned at high resolution at a rate of about 1.3 slides per minute using the newly developed digitization system. It is estimated that two conventional slide scanners would take approximately 2.5 years for the same set of slides, and this is without capturing the slide frame.
Keywords: imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • retina