Polyps Detection Using Ultralytics YOLOv8

Abstract

Several diseases can affect the digestive human system. Among them, the colorectal cancer (CRC) is the great common cause of cancer-related death in the world. The CRC is mainly caused by polyps that can grow on the inner walls of the intestines or the rectum. To improve chances of survival, prior detection, prognosis and timely treatment are crucial factors (it is about the surgical removal of polyps). The using of developed Computer-aided diagnosis systems (CADx) equipped with appropriate machine learning algorithms, and specifically deep learning methods contribute to assist patricians to obtain very pertinent detection of abnormalities in intern human body exploration. In this respect, this article exposes a deep learning model for automated polyp detection by using the You Only Look Once (YOLO) model. Earlier versions of the YOLO family have been experimented with improved performances; the present paper concerns YOLOv8 model application. This paper presents an implementation of a detection system based on the Yolo v8n model respecting the characteristics of simplicity, performance, cost and which can provide substantial help to practitioners and patients, when it is applied in the field of polyp’s detection.The real-time detection model is used with both Kvasir-SEG and BKAI-IGH NeoPolyp-Small datasets. The obtained results are compared to those obtained by the use of the YOLOv7 and show improved performances.