DETECTING LEOPARD SKIN IN THE CERVIX BY IMAGE PROCESSING METHOD

Abstract

Leopard skin appeared on the surface of the cervix in vaginitis is associated with trichomoniasis infection. Trichomoniasis is the most common sexually transmitted infections (STIs) worldwide with more than 250 new cases each year and about 70% of cases without any signs or symptoms. The appearance of leopard skin in the cervix is the origin of cervicitis and some other dangerous gynecological diseases. Without early detection and prompt treatment, the disease can become one of the causes of infertility and cervical cancer. In clinical, colposcopy combined with the use of Lugol’s iodine is a common method for detecting the presence of leopard skin. This method is considered to be of low accuracy, the diagnosis result depends on the quality of Lugol, the smear technique, and the experience of the doctor reading result. In this study, based on the distribution characteristics of blood and glycogen, we developed an automated method of detecting the presence of leopard skin from images obtained by digital colposcopes. Results of image segmentation are compared and evaluated with the traditional method.