Children born prematurely may suffer from an eye disease called retinopathy of prematurity (ROP). To estimate the severity of this disease, physicians need to type, among other things, the extent of the disease on images that are often of poor quality. The extent of ROP is measured from the optical disc. Therefore, it is essential to have an automatic method that locates and segments the optical disc. In order to contribute to a computational method that detects the optic disc in children’s pathological images, a fast-processing method is presented in this work. This method creates a template-based local binary pattern histogram. Next, the method proposes recognizing candidate windows as an optical disk from regional maxima. Then, the template is used to choose the correct optic disk. This method used thirty images from the ROPFI dataset that contains infant pathological images. The optic disk has been manually labeled. The test identifying the optic disk achieved a sensitivity of 0.95.