癌症是导致死亡的主要原因之一，其特点是健康细胞逐渐转化为癌细胞的多阶段过程。早期发现该疾病可以显著提高生存的可能性。组织学是一种将感兴趣的组织首先从患者身上切除并切成薄片的手术过程。然后，病理学家会将这些薄片放在玻璃载玻片上，用专用染料（如苏木精和伊红）对其染色，然后在显微镜下观察这些玻璃载玻片。然而，对乳腺癌活检组织病理学图像进行手动分析非常耗时。文献表明，基于深度学习算法和人工智能的自动化技术可以用于提高从乳腺癌患者获得的组织病理标本中异常的检测速度和准确性。本文重点介绍了一些关于这些算法的最新研究，并提供了各种深度学习方法的比较研究。本研究使用了乳腺癌组织病理数据库（BreakHis）。对这些图像进行处理以增强其内在特征，并对算法的准确性进行评估。分析这些图像时使用了三种卷积神经网络（CNN）模型，即视觉几何组（VGG19）模型、密集连接卷积网络（DenseNet201）模型和残差神经网络（ResNet50V2）模型。其中，DenseNet201模型表现优于其他模型，达到了91.3%的准确率。本文还包括了一些基于机器学习方法的不同分类技术的综述，其中包括基于CNN的模型，其中一些技术可能取代手动乳腺癌诊断和检测。

Cancer, a leading cause of mortality, is distinguished by the multi-stage conversion of healthy cells into cancer cells. Discovery of the disease early can significantly enhance the possibility of survival. Histology is a procedure where the tissue of interest is first surgically removed from a patient and cut into thin slices. A pathologist will then mount these slices on glass slides, stain them with specialized dyes like hematoxylin and eosin (H&E), and then inspect the slides under a microscope. Unfortunately, a manual analysis of histopathology images during breast cancer biopsy is time consuming. Literature suggests that automated techniques based on deep learning algorithms with artificial intelligence can be used to increase the speed and accuracy of detection of abnormalities within the histopathological specimens obtained from breast cancer patients. This paper highlights some recent work on such algorithms, a comparative study on various deep learning methods is provided. For the present study the breast cancer histopathological database (BreakHis) is used. These images are processed to enhance the inherent features, classified and an evaluation is carried out regarding the accuracy of the algorithm. Three convolutional neural network (CNN) models, visual geometry group (VGG19), densely connected convolutional networks (DenseNet201), and residual neural network (ResNet50V2), were employed while analyzing the images. Of these the DenseNet201 model performed better than other models and attained an accuracy of 91.3%. The paper includes a review of different classification techniques based on machine learning methods including CNN-based models and some of which may replace manual breast cancer diagnosis and detection.