Classifying Visual Field Loss in Glaucoma through Baseline Matching of Stable Reference Sequences

Shuanghui Meng
Department of Computing Curtin University of Technology Perth, Australia

Andrew Turpin
Department of Computer Science and Information Technology, RMIT University, Melbourne, 3052, Australia.

Mihai Lazarescu
Department of Computing Curtin University of Technology Perth, Australia

Jim Ivins
Department of Computing Curtin University of Technology Perth, Australia

Status

Proc Int. Conf. on Machine Learning and Cybernetics (ICMLC2005)

Abstract

Glaucoma is a common disease of the eye that often results in partial blindness. The main symptom of glaucoma is progressive loss of sight in the visual field over time. The clinical management of glaucoma involves monitoring the progress of the disease using a sequence of regular visual field tests. However, there is currently no universally accepted standard method for classifying changes in the visual field test data. Sequence matching techniques typically rely on similarity measures. However, visual field measurements are very noisy, particularly in people with glaucoma. It is therefore difficult to establish a reference data set including both stable and progressive visual fields. This paper proposes a method that uses a baseline computed from a query sequence, to match stable sequences in a database of visual field measurements collected from volunteers. The purpose of the new method is to classify a given query sequence as being stable or progressive. The results suggest that the new method gives a significant improvement in accuracy for identifying progressive sequenses, though there is a small penalty for stable sequences.