Scholarlink

In medical science, the detection of malignancies is essential for the diagnosis of cancer. The research and identification of disease symptoms through microscopic tissue inspection is known as histopathology. Pathologies carefully analyze these intricate histopathological images (HIs). The different results cause the pathologists to become subjective, and they mostly rely on the examiner’s experience. In order to categorize biopsy samples as benign or malignant, many characteristics such as color, size, form, and layout of the cells and tissues are studied. These take a lot of time and are vulnerable to the pathologist’s subjectivity. Computer-assisted analysis is required to get around this. This study examines the effectiveness of several machine learning approaches on the BreaKHis dataset, which contains of microscopic images of breast tumor tissue that were obtained with various magnification factors. Machine learning can function with little in the way of computing power. In order to categorize certain HIs of breast cancer as either malignant or benign, an evaluation of the performance of several machine learning algorithms, as well as K- closest neighbor (KNN), support vector machine (SVM), decision tree, and Gaussian naïve bayes, is being conducted in this research. When the data are scaled and normalized, KNN attains 92.62% efficiency whereas decision trees perform better through 87.98% accuracy.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An Analytical Study on Classification of Histopathological Images for Cancer Diagnosis and Progression Using Machine Learning Techniques

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Nam Vasundhara,
  • Guntoju Kalpana Devi,
  • A. Srinivasula Reddy,
  • X. S. Asha Shiny,
  • Manyala Naga Sailaja,
  • T. Upender

摘要

In medical science, the detection of malignancies is essential for the diagnosis of cancer. The research and identification of disease symptoms through microscopic tissue inspection is known as histopathology. Pathologies carefully analyze these intricate histopathological images (HIs). The different results cause the pathologists to become subjective, and they mostly rely on the examiner’s experience. In order to categorize biopsy samples as benign or malignant, many characteristics such as color, size, form, and layout of the cells and tissues are studied. These take a lot of time and are vulnerable to the pathologist’s subjectivity. Computer-assisted analysis is required to get around this. This study examines the effectiveness of several machine learning approaches on the BreaKHis dataset, which contains of microscopic images of breast tumor tissue that were obtained with various magnification factors. Machine learning can function with little in the way of computing power. In order to categorize certain HIs of breast cancer as either malignant or benign, an evaluation of the performance of several machine learning algorithms, as well as K- closest neighbor (KNN), support vector machine (SVM), decision tree, and Gaussian naïve bayes, is being conducted in this research. When the data are scaled and normalized, KNN attains 92.62% efficiency whereas decision trees perform better through 87.98% accuracy.