-
Book Overview & Buying
-
Table Of Contents
-
Feedback & Rating

Computer Vision on AWS
By :

When humans view visual information, we instinctively detect labels for objects, scenes, and activities. Amazon Rekognition offers similar capabilities as easy-to-use APIs that don’t require machine learning expertise. Using the Rekognition management console, you learned how to upload messages and view the response payloads. Next, you programmatically repeated that process using the Python boto3
module. Additionally, this chapter introduced the PIL for drawing bounding boxes on the example images.
Amazon Rekognition’s built-in label detection supports over 2,500 labels with more than 250 pieces of supporting bounding box information. While this breadth covers numerous use cases, it won’t cover every business-specific need, such as finding Packt Publishing’s logo. In that case, you’ll need to train Amazon Rekognition Custom Labels. Join us in the next chapter to learn how to use this feature without requiring undifferentiated heavy lifting...