Packt Publishing – Advanced Computer Vision with TensorFlow | 543.26 MB
This video course is a practical guide to implementing TensorFlow in production and is packed with step-by-step instructions, working examples, and helpful advice about building your neural networks with TensorFlow, where you’ll learn to build separable convolutional neural networks. This practical course is divided into clear byte-size chunks so you can learn at your own pace and focus on the areas that interest you the most.
TensorFlow has been gaining immense popularity over the past few months, due to its power and simplicity to use. This video will help you leverage the power of TensorFlow to perform advanced image processing. This course is a continuation of the Intro to Computer Vision course, building on top of the skills learned in that course. In this course, you’ll dive deeper as we cover more advanced computer vision concepts.
You will implement multiple state-of-the-art deep learning papers from scratch using the TensorFlow-Keras API. This course will teach you how to construct efficient CNN architectures with CNN Squeeze layers and delayed downsampling . You’ll learn about residual learning with skip connections and deep residual blocks, and see how to implement a deep residual neural network for image recognition. You’ll find out about Google’s Inception module and depthwise separable convolutions and understand how to construct an extreme Inception architecture with TF-Keras.
Finally, you’ll be introduced to the exciting new world of adversarial neural networks, which are responsible for recent breakthroughs in synthetic image generation and implement an auxiliary conditional GAN.
What You Will Learn
• Build efficient architecture for convolutional neural networks
• Construct a residual learning neural network for image recognition
• Build depthwise separable convolutional neural networks
• Construct conditional Generative Adversarial Networks (GAN)
• Build an advanced and powerful multi-class image classifier
• Build functional model class and methods with TensorFlow-Keras’ Functional API
• Build a computational graph representation of a Neural Network from state-of-the-art deep
• Optimize a neural network with stochastic gradient descent and other advanced optimization