|
所在平台: Udemy |
课程主页: https://www.udemy.com/course/data-science-cnn-opencv-breast-cancer-detection/
课程评论：没有评论
本课程名为“数据科学：CNN与OpenCV：乳腺癌检测”，旨在教授学员如何利用AI和机器学习算法，通过全切片图像（包括浸润性导管癌IDC阳性和阴性病例）来检测乳腺癌。 课程特色： * **实战项目：** 这是一个实践性强的项目，将一步步指导学员使用TensorFlow、CNN、OpenCV和Python创建和评估深度学习模型。 * **内容全面：** 涵盖了从数据探索、数据增强、数据生成器、定制预训练模型（如ResNet50）到从头构建CNN模型架构、模型检查点、模型构建与评估，并最终利用训练好的模型进行乳腺癌检测的全流程。 * **技术栈：** 重点使用TensorFlow、CNN、OpenCV和Python。 * **学习平台：** 主要在Google Colab环境中进行教学。 * **学习目标：** 帮助学员掌握构建高精度乳腺癌细胞检测模型的技术。 * **课程产出：** 学员将获得AutomationGig的结业证书，并提供Jupyter Notebook及其他项目文件。 课程主要任务分解： 课程被细分为47个任务，涵盖了项目概览、Google Colab使用、项目文件夹结构、数据集理解、项目环境设置、配置文件、库导入、数据可视化（数据分布、样本展示）、数据处理方法（文件计数、绘制训练/验证曲线）、类别权重计算、数据增强技术、数据生成器实现、CNN原理、OpenCV介绍、预训练模型（ResNet50）讲解、CNN关键层（Conv2D, MaxPooling2D等）解释、使用ResNet50和自定义CNN构建模型、模型编译（优化器、损失函数）、模型检查点应用、模型训练（Epoch, Batch Size）、模型预测、评估指标（分类报告、混淆矩阵、AUC-ROC）计算和可视化、模型序列化（保存/加载）、以及如何进一步提升模型性能。 总结： 该课程提供了一个从数据加载到云端预测的完整乳腺癌图像分类模型构建流程，旨在让学员在较短时间内深入理解深度学习，并能独立完成乳腺癌检测模型的开发。课程强调实践操作，通过具体代码演示，帮助学员掌握现代深度学习在医疗诊断领域的应用。
If you want to learn the process to detect whether a person is suffering breast cancer using whole mount slide images of positive and negative Invasive Ductal Carcinoma (IDC) with the help of AI and Machine Learning algorithms then this course is for you.In this course I will cover, how to build a model to predict whether a patch of a slide image shows presence of breast cancer cells with very high accuracy using Deep Learning Models. This is a hands on project where I will teach you the step by step process in creating and evaluating a deep learning model using Tensorflow, CNN, OpenCV and Python.This course will walk you through the initial data exploration and understanding, Data Augumentation, Data Generators, customizing pretrained Models like ResNet50 and at the same time creating a CNN model architecture from scratch, Model Checkpoints, model building and evaluation. Then using the trained model to detect the presence of breast cancer.I have split and segregated the entire course in Tasks below, for ease of understanding of what will be covered.Task 1 : Project Overview.Task 2 : Introduction to Google Colab.Task 3 : Understanding the project folder structure.Task 4 : Understanding the dataset and the folder structure.Task 5 : Setting up the project in Google Colab_Part 1Task 6 : Setting up the project in Google Colab_Part 2Task 7 : About Config and Create_Dataset FileTask 8 : Importing the Libraries.Task 9 : Plotting the count of data against each class in each directoryTask 10 : Plotting some samples from both the classesTask 11: Creating a common method to get the number of files from a directoryTask 12: Defining a method to plot training and validation accuracy and lossTask 13: Calculating the class weights in train directoryTask 14: About Data Augmentation.Task 15: Implementing Data Augmentation techniques.Task 16: About Data Generators.Task 17: Implementing Data Generators.Task 18: About Convolutional Neural Network (CNN).Task 19: About OpenCV.Task 20: Understanding pre-trained models.Task 21: About ResNet50 model.Task 22: Understanding Conv2D, Filters, Relu activation, Batch Normalization, MaxPooling2D, Dropout, Flatten, DenseTask 23: Model Building using ResNet50Task 24: Building a custom CNN network architecture.Task 25: Role of Optimizer in Deep Learning.Task 26: About Adam Optimizer.Task 27: About binary cross entropy loss function.Task 28: Compiling the ResNet50 modelTask 29: Compiling the Custom CNN ModelTask 30: About Model CheckpointTask 31: Implementing Model CheckpointTask 32: About Epoch and Batch Size.Task 33: Model Fitting of ResNet50, Custom CNNTask 34: Predicting on the test data using both ResNet50 and Custom CNN ModelTask 35: About Classification Report.Task 36: Classification Report in action for both ResNet50 and Custom CNN Model.Task 37: About Confusion Matrix.Task 38: Computing the confusion matrix and and using the same to derive the accuracy, sensitivity and specificity.Task 39: About AUC-ROCTask 40: Computing the AUC-ROCTask 41: Plot training and validation accuracy and lossTask 42: Serialize/Writing the model to diskTask 43: Loading the ResNet50 model from driveTask 44: Loading an image and predicting using the model whether the person has malignant cancer.Task 45: Loading the custom CNN model from driveTask 46: Loading an image and predicting using the model whether the person has malignant cancer.Task 47: What you can do next to increase model's prediction capabilities.Machine learning has a phenomenal range of applications, including in health and diagnostics. This course will explain the complete pipeline from loading data to predicting results on cloud, and it will explain how to build an Breast Cancer image classification model from scratch to predict whether a patch of a slide image shows presence of Invasive Ductal Carcinoma (IDC).Take the course now, and have a much stronger grasp of Deep learning in just a few hours!You will receive:1. Certificate of completion from AutomationGig.2. The Jupyter notebook and other project files are provided at the end of the course in the resource section.So what are you waiting for?Grab a cup of coffee, click on the ENROLL NOW Button and start learning the most demanded skill of the 21st century. We'll see you inside the course!Happy Learning!![Please note that this course and its related contents are for educational purpose only][Music: bensound]
