Machine learning with different digital images classification in laparoscopic surgery

The evaluation of the effectiveness automatic computer diagnostic (ACD) systems developed based on two classifiers – HAAR features cascade and AdaBoost for laparoscopic diagnostics appendicitis ovarian cysts in women with chronic pelvic pain is presented. training cascade, were performed images/ frames, which have been extracted from video gained diagnostics. Both gamma-corrected RGB converted into HSV frames used training. Descriptors images method Local Binary Pattern (LBP), includes both data color characteristics («modified LBP» - MCLBP) textural characteristics, later classifier Classification test revealed that highest recall was achieved after MCLBP descriptors 0.708, case 0.886.
 Developed AdaBoost-based ACD system a 73.6% correct classification rate (accuracy) 85.4% cysts. accuracy identification 0.653 (RGB) 0.708 (HSV) values. It concluded feature-based turned to be less effective when compared trained descriptors. Ovarian better diagnosed ACD.