Pterygium is the tissue growth on the cornea of the eye. It can cause vision obstacle if the condition of pterygium is severe. In the conventional method for identify the presence of pterygium is by manual observing and using the fluorescein dyes staining for enhancement. In this work, the observing of the pterygium is proposed using the computerized automation classification. For this purpose, two groups of images are normal and pterygium is used. These images are undergoing the pre-processed to gain ROI in corneal of the eye. The techniques used such as, contrast-limited adaptive histogram equalization (CLAHE), image center, morphological, find centroid, and the segmentation. The segmentation images are then; classify using the bags of the features (BoF) model. The processes involved such as features extraction using the speeded uprobust features (SURF), K-mean sclustering and using support vector machine (SVM) for classification. A total of 100 samples of images used, 70% of it used for training and the remaining for validation. The classification result obtained from the proposed system is effective to tune of 98% of the overall accuracy. This can contribute to improved performance in the pterygium classification by using the proposed classification system. In order to be used in clinics and hospitals, studies need to be done by increasing the number of samples used to obtain more reliable results.
