A photograph of a fruit stall in a market with a woman buying fruit. Items in the image have a tag label next to them with the fruit or object name. This symbolizes how the SageMaker image classification algorithm works.

Image Classification Algorithm

The SageMaker Image Classification algorithm can apply multiple labels to an image depending on what objects are identified. Objects are either identified, or not, there are no probability scores.

Attributes

Problem attributeDescription
Data types and formatImage
Learning paradigm or domainImage Processing, Supervised
Problem typeImage and multi-label classification
Use case examplesLabel/tag an image based on the content of the image

Training

The recommended format for training data for the image classification algorithm is the Apache MxNet recordIO format, although it will also accept jpeg and png. Incremental training can be used to speed up training. This is where a previously trained model is used for the training data.

Model artifacts and inference

DescriptionArtifacts
Learning paradigmSupervised
Request formatapplication/x-image, jpeg, png
Resultapplication/x-image

Processing environment

Only GPU instances are recommended for the Image Classification algorithm. Multiple GPU instances and single instances with multiple GPUs can be used.

AWS Innovate | Intro to Deep Learning: Building an Image Classifier on Amazon SageMaker

This is a one hour and four minute video featuring the Image Classification algorithm, by Gabe Hollombe, from AWS.

Credits

Woman and vegetable store photo by Clem Onojeghuo on Unsplash

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *