Examine how to work with Convolutional Neural Networks, and discover how to leverage TensorFlow to build custom CNN models for working with images.
Learning Objectives
TensorFlow: Convolutional Neural Networks for Image Classification
- Course Overview
- compare the working of the visual cortex with a neural network
- apply convolution to an input matrix and generate a result
- use scikit-image to read in an image
- instantiate a convolutional kernel to use with a convolutional layer
- work with convolutional layers to detect edges in the input image
- recognize how pooling works and its use in a convolutional neural network
- recognize how hyperparameters are used to design the convolutional neural network
- identify the standard structure of a convolutional neural network
- define an overfitted model and the bias-variance trade-off
- identify regularization, cross-validation, and dropout as ways to mitigate overfitting
- describe how to use the CIFAR-10 dataset for image classification
- demonstrate how to split the dataset into training and test images
- create placeholders and variables for the convolutional neural network
- define convolutional and pooling layers programmatically
- demonstrate how to run training and prediction on the CIFAR-10 dataset
- define the role of convolutional and pooling layers in a convolutional neural network
Framework Connections
The materials within this course focus on the NICE Framework Task, Knowledge, and Skill statements identified within the indicated NICE Framework component(s):
Specialty Areas
- Software Development
Feedback
If you would like to provide feedback for this course, please e-mail the NICCS SO at NICCS@hq.dhs.gov.