Recognition of Diabetic Retina Patterns using Machine Learning

Preview this chapter:

Recognition of Diabetic Retina Patterns using Machine Learning, Page 1 of 1

< Previous page | Next page > /docserver/preview/fulltext/9789815238990/chapter-6-1.gif

Medical images contain data related to the diseases and it should be interpreted accurately. However, its visual interpretation is quite complex/timeconsuming and only medical experts can examine this data precisely. In case of diabetes, the retina may be damaged and it is quite complex to examine its impact on the retina because there are a lot of vessels inside the human eyes that may be changed due to this disease and manual interpretation of these changes consumes excessive time. In order to overcome this issue, in this paper, a contour-based pattern recognition method (CBPR) is introduced that can recognize multiple patterns in sample retina images. Comparative analysis with the segmentation-based method (SBPR) shows that it outperforms in terms of performance parameters (i.e. Accuracy/Sensitivity/ Specificity etc.).&nbsp;