[论文解读] Brain Tumor Detection and Classification based on Hybrid Ensemble Classifier

Ginni Garg, Ritu Garg|arXiv (Cornell University)|Jan 1, 2021

Brain Tumor Detection and Classification参考文献 20被引用 46

一句话总结

该论文提出一种混合集成分类器（KNN-RF-DT 通过多数投票）用于脑肿瘤检测以及从 MRI 的良性/恶性分类，使用 Otsu 分割和来自 SWT、PCA、GLCM 的特征，在2556张图像上实现 97.305% 的准确率。

ABSTRACT

To improve patient survival and treatment outcomes, early diagnosis of brain tumors is an essential task. It is a difficult task to evaluate the magnetic resonance imaging (MRI) images manually. Thus, there is a need for digital methods for tumor diagnosis with better accuracy. However, it is still a very challenging task in assessing their shape, volume, boundaries, tumor detection, size, segmentation, and classification. In this proposed work, we propose a hybrid ensemble method using Random Forest (RF), K-Nearest Neighbour, and Decision Tree (DT) (KNN-RF-DT) based on Majority Voting Method. It aims to calculate the area of the tumor region and classify brain tumors as benign and malignant. In the beginning, segmentation is done by using Otsu's Threshold method. Feature Extraction is done by using Stationary Wavelet Transform (SWT), Principle Component Analysis (PCA), and Gray Level Co-occurrence Matrix (GLCM), which gives thirteen features for classification. The classification is done by hybrid ensemble classifier (KNN-RF-DT) based on the Majority Voting method. Overall it aimed at improving the performance by traditional classifiers instead of going to deep learning. Traditional classifiers have an advantage over deep learning algorithms because they require small datasets for training and have low computational time complexity, low cost to the users, and can be easily adopted by less skilled people. Overall, our proposed method is tested upon dataset of 2556 images, which are used in 85:15 for training and testing respectively and gives good accuracy of 97.305%.

研究动机与目标

• 及早且准确的脑肿瘤诊断以改善患者结局。
• 开发一种在相对较小数据集上也能良好工作的非深度学习方法，且计算效率高。
• 将分割、特征提取和混合集成分类器结合起来，将肿瘤分类为良性或恶性。
• 在脑部 MRI 图像数据集上评估所提方法，并将其性能与传统分类器进行比较。

提出的方法

• 对肿瘤区域进行 Otsu 阈值分割。
• 利用 Stationary Wavelet Transform (SWT)、Principal Component Analysis (PCA) 和 Gray Level Co-occurrence Matrix (GLCM) 提取特征，以获得十三个特征。
• 使用具有 Majority Voting 判定规则的混合集成分类器（KNN-RF-DT）进行分类。
• 从分割中确定肿瘤面积，并基于集成输出将其分类为良性或恶性。
• 强调传统分类器相较于深度学习的优势，如较小的训练数据需求和较低的计算成本。

实验结果

研究问题

• RQ1 在 MRI 数据上，基于多数投票的 KNN、Random Forest、Decision Tree 的混合集成是否能实现高准确率的脑肿瘤检测与分类？
• RQ2 通过 Otsu 方法的分割结合 SWT、PCA、GLCM 特征，是否能对肿瘤区域与正常组织提供鲁棒的辨别？
• RQ3 与传统分类器和深度学习方法相比，该方法在数据需求和计算方面表现如何？
• RQ4 在测试的 MRI 数据集上的总体准确率是多少，该方法在 85:15 的训练-测试划分下的可靠性如何？

主要发现

• 在一个包含 2556 张 MRI 图像、85:15 训练-测试划分的数据集上实现 97.305% 的准确率。
• 使用 Otsu 分割再进行特征提取（SWT、PCA、GLCM）得到用于分类的十三个特征。
• 混合集成 KNN-RF-DT 以多数投票在本研究中相较于传统分类器提升了性能。
• 作者认为传统分类器在训练数据需求更少、计算成本较低等方面相对于深度学习具有优势。