Safe autonomy (ECE498SM) Lecture 5: Image classification

views comments

•Image classification: training with labels, testing

•Training and testing both require features

•Bag of visual words / vocabulary as a feature

•Learning visual vocabulary using clustering to deal with quantization

•Capturing local spatial ordering with pyramids

•We will study k-means clustering later

•Classification

•Nearest neighbor classification

•Hyperparameters for K-NN (distance measure, k)

•Best practices for classification

	Contact Technology Services to report an issue, offer feedback, or request assistance.
	Technology Services Home \| Kaltura Help \| Privacy Policy
	Illinois Media Space, © 2022 Board of Trustees of the University of Illinois

Related Media