Visão Geral
Curso Medical Image Analysis with Deep Learning. Embarque em uma jornada transformadora no domínio da análise de imagens médicas com cursos superiores. Este Curso Medical Image Analysis with Deep Learning abrangente oferece uma exploração aprofundada de várias modalidades de imagens médicas, incluindo raios X, ressonância magnética e imagens ultrassônicas. Com sessões práticas sobre operações básicas de imagens usando Python, você compreenderá as complexidades da textura em imagens médicas e dominará a arte da extração clássica de recursos.
Mas o aprendizado não termina aqui. Aprofunde-se no mundo das redes neurais, entendendo as complexidades dos Autoencoders, tanto esparsos quanto sem ruído. Além disso, aproveite o potencial do Deep Learning com Redes Neurais Convolucionais (CNN) e suas vastas aplicações na análise de imagens médicas. Com uma mistura de conhecimento teórico e experiência prática, este curso é feito sob medida para aqueles que desejam revolucionar o diagnóstico de saúde e o planejamento de tratamento por meio do poder do aprendizado profundo.
Conteúdo Programatico
Introduction to Medical Images
- X-ray and CT Imaging
- Magnetic Resonance Imaging
- Ultrasound Imaging
- Optical Microscopy and Molecular Imaging
Basic Image operation with Python
- Read and Display Image
- Covert Colour Image to Grayscale Image
- Cropping and Resizing an Image
- Rotating Image
- Histogram Equalization
- Blurring an Image
Texture in Medical Images
- Texture characterization – Statistical vs Structural
- Co-occurrence Matrix
- Orientation Histogram
- Local Binary Pattern (LBP)
- Texture from Fourier features
- Wavelets
- Feature extractions for Image (Medical/General)
Neural Network for Visual Computing
- Simple Neuron
- Neural Network formulation
- Learning with Error Propagation
- Gradient Checking and Optimization
Deep Learning
- What is Deep Learning?
- Families of Deep Learning
- Multilayer Perceptron
- Learning Rule
- Autoencoders
- Retinal Vessel Detection using Autoencoders
Stacked, Sparse, Denoising Autoencoders
- Stacking Autoencoders
- Ladder wise pre-training and End-to-End Pre-training
- Denoising and Sparse Autoencoders
- Ladder Training
- End-to-End Training
- Medical Image classification with Stacked Autoencoders
Convolutional Neural Network (ConvNet)
- What is ConvNet?
- Difference between Fully connected NN and ConvNet.
- Stride, Padding, and Pooling
- Deconvolution
- ReLU Transfer Function
Image Classification with CNN
- Convolutional Autoencoder
- LeNet for Image Classification
- AlexNet for Image Classification
Improving Deep Neural Network
- Batch Normalization, Dropout
- Tuning Hyper-parameters to improve performance of NN.
- Learning Rate Annealing
- Different Cost Functions
Deep CNN and its application to Medical Images
- Vgg16, ResNet34, GooleNet, and DenseNet121
- Transfer Learning
- Pneumonia detection from Chest X-rays with Deep CNN.
- White blood cell classification with CNN
Object Localization
- Activation pooling for object localization
- Region proposal Network
- Sematic segmentation
- UNet
- Retinopathy Image segmentation with UNet
Spatio-Temporal Deep Learning
- Understanding Video analysis
- Recurrent Neural Network (RNN)
- Long Short Term Memory (LSTM)
- Activity recognition using 3D-CNN
- Analysis of Brain Images