| |
| ATLAS 9000 Corneal Topographer |
| |
PathFinder II Corneal Analysis software
One of the most interesting and useful aspects of the ATLAS 9000 corneal topographer is the PathFinder II Corneal Analysis software. PathFinder II compares the measured corneal topography to an extensive clinical database and provides a probability that the topography exam falls into specific categories. This is an extremely useful aid to interpretation, but how does it work?
In mathematical terms, PathFinder II incorporates a Support Vector Machine (SVM) algorithm that uses a clinical database to perform classification by constructing “multi-dimensional hyperplanes” to optimally separate the various corneal conditions to which it is trained. In other words, the SVM looks for common patterns in corneal topography maps compared to those in the clinical database and rates the likelihood it falls within predefined clinical categories.
SVM’s have received widespread recognition for their ability to discover hidden relationships in complex datasets and have been implemented in numerous applications including image analysis and decryption. PathFinder II evaluates 12 different corneal parameters to classify the cornea into five categories: normal, suspect keratoconus, post-myopic laser vision correction, post-hyperopic laser vision correction or “other” corneal conditions.
The clinical database is made up of 369 eyes of 332 subjects and is made up of both normal and “abnormal” corneas. Thus it is referred to as a clinical database rather and a normative database.
Download a ATLAS 9000 (2,395kB) brochure
Top
| |
 |
 |
Visit us at both the:
SRC | 16th – 18th May & ODMAFair | 10th – 12th July
to find out about our conference deals including special interest rates in conjunction with Investec Experien. In particular, ask us about our specials on our OCT technology; it might be more affordable than you think. |
 |
 |
 |
 |
|
Editor: Matthew Wensor
Product Manager
Ophthalmic Systems
Medical Business Group
Email Us |
| Ph: 02 9020 1333 |
 |
 |
 |
 |
 |
|
|