Classification of Cataract Slit-Lamp Image Based on Machine Learning

Cataracts are diseases caused by the presence of proteins in the lens that form abnormal and gradually enlarged clumps that will interfere with vision by blocking the light entering through the lens. Identification of cataracts is done by taking the image of the eye with a slit-lamp tool from the front of the eye. Slit-lamp images can provide information about the condition of the pupils that can only be analyzed by the doctor manually based on doctor's observation and doctor's experience that can cause different analysis in determining the actual eye condition. Things that are considered by the doctor in analyzing cataracts are the level of opacity in the eyes and the area covered by the turbid. Identification and classification with slit-lamp images can be performed better and more accurately using image processing techniques. Firstly, the grayscale method, median filter method and canny method is used to preprocess the slit-lamp images. Next, the hough circular method is used to automatically segment pupil from slit-lamp images. After the segmentation process, we use pixel scanning to extract mean intensity and uniformity from the pupil image. After the feature extraction process, classification is done by single perceptron based on the extracted feature. This research is expected to help the doctor to do cataracts classification so that the classification process will be easier and more accurate. Based on the test result show that the accuracy of the system is 96.6%.