Computerized Detection of Macular Edema Using OCT Images Based on Fractal Texture Analysis

Macular edema (ME) is a significant cause that results in blindness among majority. It happens due to the anomalous leakage and builds up of fluids within macular region. Anything that relates with the performance of eyes can drive the chances to become a ME patient. ME can occur as a downstream of eye related surgeries, degradation due to aging or any eye disorders causing inflammation. The main fact about ME that if the disease is not identified and diagnosed at the earlier stages, the chance of recovery is minimal and can ultimately affect the ability to see. One of the main causes of ME is the disease which can injure blood vessels in the retina. Laser photocoagulation and vitrectomy are the common method available now for diagnosing the disease. Optical Coherence Tomography alias OCT, an advanced imaging technique to capture retinal layer region. Different algorithms were implemented to detect ME from OCT images, but early detection of ME is not possible. This paper utilizes segmentation based fractal texture analysis (SFTA) to derive the feature vector. Graph based segmentation employs in the detection of layers and QDA classifies the ME images. This algorithm will help the ophthalmologist to treat the patient at early stages. The algorithm is deployed successfully on a macular edema dataset, with 97.5% accuracy rate.