Predictive Diagnosis of Glaucoma Based on Analysis of Focal Notching along the Neuro-Retinal Rim Using Machine Learning

Automatic detection of glaucoma via fundus imaging and artificial intelligence: A review

“…They compared this framework with the same methodology but using only the CDR feature yet, they found this resulted in significantly lower performance metrics. Thus, indicating the relevance of the other parameters used; however, this is to be further examined to test the generalizability of the other features for glaucoma classification with external datasets ( 65 ). Similarly, Khalil and coworkers( 52 ) found an improved performance by combining structural and textural features for classification ( Table 3 ).…”

“…From the four papers that used support vector machine classifiers with linear kernels, Narasimhan and Vijayarekha only provided the metric of accuracy which was 95% ( 69 ). Mukherjee and coworkers obtained an accuracy of 87%, sensitivity of 86.4% and specificity of 90% ( 65 ). More recently, Pathan and coworkers achieved an accuracy of 96.66%, sensitivity of 100% and specificity of 95% with the publicly available DRISHTI database but on external testing with a private database, this reduced to an accuracy of 90%, sensitivity of 93.47% and specificity of 91.2% ( 59 ).…”

“…The four quadrants of the NRR are typically expected to satisfy the inferior-superior-nasal-temporal (ISNT) rule (I > S > N > T) ( 65 ). Whilst the cup-to-disc ratio parameter focuses on the optic cup size with respect to the optic disc, the ISNT rule focuses on the NRR width i.e., the area between the boundary of the optic cup and disc ( 65 ).…”

“…The four quadrants of the NRR are typically expected to satisfy the inferior-superior-nasal-temporal (ISNT) rule (I > S > N > T) ( 65 ). Whilst the cup-to-disc ratio parameter focuses on the optic cup size with respect to the optic disc, the ISNT rule focuses on the NRR width i.e., the area between the boundary of the optic cup and disc ( 65 ). The ISNT rule follows that the inferior rim is thicker than the superior rim, which is thicker than the nasal rim, which is thicker than the temporal rim in a healthy eye ( 24 ).…”

Improving Early Breast Cancer Detection with Artificial Intelligence: A Promising Approach

“…ML algorithms are used to store and process large datasets, which are then used to train predictive models and interpret generalizations. 6,7 DL, on the other hand, utilizes advanced neural networks to identify complex patterns in medical images that are difficult for human to recognize. [8][9][10] It has also been hypothesized that DL-based tumor segmentation exhibit better and accurate prediction of triple negative breast cancer minimizing manual interventions.…”