Detection of Colorectal Polyps from Colonoscopy Using Machine Learning: A Survey on Modern Techniques

Abstract

Given the increased interest in utilizing artificial intelligence as an assistive tool in the medical sector, colorectal polyp detection and classification using deep learning techniques has been an active area of research in recent years. The motivation for researching this topic is that physicians miss polyps from time to time due to fatigue and lack of experience carrying out the procedure. Unidentified polyps can cause further complications and ultimately lead to colorectal cancer (CRC), one of the leading causes of cancer mortality. Although various techniques have been presented recently, several key issues, such as the lack of enough training data, white light reflection, and blur affect the performance of such methods. This paper presents a survey on recently proposed methods for detecting polyps from colonoscopy. The survey covers benchmark dataset analysis, evaluation metrics, common challenges, standard methods of building polyp detectors and a review of the latest work in the literature. We conclude this paper by providing a precise analysis of the gaps and trends discovered in the reviewed literature for future work.

UY-NET: A Two-Stage Network to Improve the Result of Detection in Colonoscopy Images

“…The table shows that the AP and IoU values for YOLOv4 and YOLOv3-spp are all above 0.81, indicating that the two YOLO models detect polyps to an adequate level. This finding aligns with what previous research has reported [34]. For instance, Doniyorjon et al [35] tested five YOLO algorithms (i.e., YOLOv3, YOLOv3-tiny, YOLOv4, YOLOv4tiny, and YOLOv4-tiny with the Inception-ResNet-A block), and all models were found to achieve at least 89% training accuracy and 85% testing accuracy.…”

“…To the best of our knowledge, this study may be the first attempt to create a two-stage network by combining U-Net and YOLOv4. As ELKarazle et al ( [34], p. 10) argued, "the YOLO architecture has been the preferred go-to solution for real-time detection tasks as it can process 45 frames per second". This feature makes it popular among researchers and one of the most used methods for polyp detection.…”

Detecção de Pólipos em Imagens de Colonoscopia utilizando Aprendizado Profundo

“…R-CNN usa redes neurais convolucionais para localizar objetos de interesse e realizar a extração de características de forma independente de cada região de interesse para processamento posterior. Alguns trabalhos apresentaram bons resultados na detecção de pólipos utilizando modelos baseados no modelo R-CNN, detectando pólipos de formas e tamanhos distintos [ELKarazle et al 2023].…”

“…A aprendizagem profunda (deep learning) é um ramo da IA que se baseia em redes neurais profundas e tem a capacidade de extrair automaticamente características relevantes em imagens médicas. Dessa forma, estudos recentes têm surgido sobre o uso de modelos de redes neurais convolucionais para detecção de pólipos em imagens de exames de colonoscopia [ELKarazle, K. et al 2023].…”

“…Detecção de pólipo colorretal no contexto da inteligência artificial significa o processo de treinar um modelo de aprendizado de máquina (machine learning) para aprender um conjunto de características específicas que representam os pólipos para então detectar eles por meio de uma caixa delimitadora (bounding box) dado uma imagem ou vídeo de entrada [ELKarazle, K. et al 2023].…”