Becoming one of the biggest national threatening blindness cause, Cataract is the 81.25 percent of the 1.6 million blindness cases in Indonesia. Where each year it increases with 250,000 new cases. The high rate of this national blindness case is not balanced with the fulfillment of a balanced and spread health facility. The ideal surgery number is supposed to reach 80 percent, while in Indonesia it only reached 45 percent. Aside from that, the inequality of available doctors in one area compared to another area makes the health care facility cannot be accessed by everyone equally. Almost half of the ophthalmologists in Indonesia live on Java island. The amount of the Cataract sufferer can be minimalized through Cataract early detection. SlitLamp is conventional medical equipment that can check the eye’s condition with the lamp’s brightness. The early check-ups are administered with a doctor’s analysis according to the output image from the Slit-Lamp check-up. The slit-lamp examination needs to be administered by professionals and special equipment, which makes the medical bill for this treatment relatively expensive. The difficulties of medical access, especially in remote areas, makes the society to be reluctant to do early eye medical check-up. The Head of the Indonesia Ophthalmologist Union (PERDAMI) hopes that by 2030 the medical treatments can be accessed by every citizen of Indonesia without any difficulties. Based on this problem, a detection system that is optimal, cheap, and can be accessed by the society is needed. This is the reason for the writer to research the cataract classification system based on mean intensity of eye image. The output classification is normal, immature cataract, and mature cataract. The amount of the Cataract sufferer can be minimalized through Cataract early detection. Based on this problem, a detection system that is optimal, cheap, and can be accessed by society is needed. This is the reason for the writer to propose this research. Panoptes is a computer vision-based cataract check platform available in online and offline modes. This system extracting features from the image using histogram gray, to acquire mean intensity. There will be a linear regression analysis process to determine the cataract classification (normal, immature cataract, and mature cataract) according to mean intensity data. The system also identifies whether the user needs further medical examination. The system determines this by applying a fuzzy logic algorithm based on the eye prediction results and the quality of the input image. Image quality assessment in this system is carried out using the BRISQUE (Blind / Referenceless Image Spatial Quality Evaluator) algorithm. This research has 90,6% accuracy, 94% F1 score, 89% precission, and 82% specificity. Compared to other existing innovations, Panoptes has some beneficial points such as: (1) Panoptes has online as well as offline mode. Online via web and offline via PC application, (2) Panoptes does not only detect but also classify cataracts, (3) Panoptes is easy to use and does not require medical experts, and (4) Panoptes has a medical advice feature to determine whether a patient needs further examination. As the conclussion that Panoptes have been proven to be able to detect and classify cataracts, Panoptes system accuracy, specificity, sensitivity/recall, precision, and F1 scores of more than 80% make Panoptes feasible to use, Panoptes can help make it easier for people to conduct early eye cataract examinations, and Panoptes can provide medical advice to users for further action and treatment.
Panoptes : Digital Platform for Early Detection and Classification of Cataract Eyes
Short Description:
Cataract becoming one of the biggest national and international threatening blindness cause. The high rate of this national blindness case in Indonesia is not balanced with the fulfillment of the health facility. The cataract surgery number in Indonesia still below the international ideal surgery number. Slit-lamp examination as the conventional method needs to be administered by professionals and special equipment, which makes the medical bill for this treatment relatively expensive. The inequality of available doctors in one area compared to another area also makes the health care facility cannot be accessed by everyone equally. The amount of the Cataract sufferer can be minimalized through Cataract early detection. Based on this problem, a detection system that is optimal, cheap, and can be accessed by society is needed. This is the reason for the writer to propose this research. Panoptes is a computer vision-based cataract check platform available in online and offline modes. This system extracting features from the image using histogram gray, to acquire mean intensity. There will be a linear regression analysis process to determine the cataract classification (normal, immature cataract, and mature cataract) according to mean intensity data. The system also identifies whether the user needs further medical examination. The system determines this by applying a fuzzy logic algorithm based on the eye prediction results and the quality of the input image. Image quality assessment in this system is carried out using the BRISQUE (Blind / Referenceless Image Spatial Quality Evaluator) algorithm. This research has 90% accuracy. With the above advantages, Panoptes can help the community in the early detection of cataracts.