- News
- Subject and Goals
- Council and Annual Reports
- Contacts and Working Groups
- Lectures and Presentations
- Conferences and Seminars
- Publications
- Application Demos
- International Travels
- Visitors
- Past Events
- E-Synergy Login
SEARCH
MEMBERS OF RESEARCH CENTER
Research Centre
Data - Algorithms - Decision Making
Data - Algorithms - Decision Making
Established in 2005 under support of MŠMT ČR (project 1M0572)
Publications
Non-rigid registration of time series of auto-fluorescent HRA 2D images
Typ:
Conference paper
Authors:
Serie:
Praha: UTIA, Czech Academy of Sciencies
Year:
2005
Pages:
3 - 3
Contact:
attachment1:
Anotation:
In the recent decades, correlation between the distribution of lipofuscin (LF) in the retinal pigment epithelium (RPE) and various ophthalmic diseases has been proven. Distribution of LF can be measured in-vivo by excitation of RPE using 488nm blue argon laser and subsequent detection of auto-fluorescence radiation of LF making use of confocal scanning laser ophthalmoscope (Heildelberg Retina Angiograph, HRA).
Unfortunately, since auto-fluorescence is of low intensity, pure HRA scans are of very poor signal/noise ratio (SNR) and can not be used for reliable determination of the LF distribution directly. Improvement of SNR is possible by enlarging photo-detector or extending the acquisition time. Nevertheless, both methods lead to worse spatial resolution. Hence, we use time series of 9 pure HRA scans and perform averaging. In respect to the eye motion, some type of alignment has to be done before averaging. This work deals with the non-rigid multi resolution registration method applicable to preserve spatial resolution while improving SNR by averaging.
Unfortunately, since auto-fluorescence is of low intensity, pure HRA scans are of very poor signal/noise ratio (SNR) and can not be used for reliable determination of the LF distribution directly. Improvement of SNR is possible by enlarging photo-detector or extending the acquisition time. Nevertheless, both methods lead to worse spatial resolution. Hence, we use time series of 9 pure HRA scans and perform averaging. In respect to the eye motion, some type of alignment has to be done before averaging. This work deals with the non-rigid multi resolution registration method applicable to preserve spatial resolution while improving SNR by averaging.