Amazon Rekognition
For image classification in the visible spectrum, models are frequently created with transfer learning and fine-tuning from a pretrained neural network. You can automate the task of network selection and training by using the Amazon Rekognition service.
Amazon Rekognition provides a standard set of classification labels. A
label is an object or concept (including scenes and
actions) found in an image or video based on its contents. For example, an image of
people on a tropical beach may contain labels such as Palm Tree
(object), Beach
(scene), Running
(action), and
Outdoors
(concept). For more information about the labels supported
by Amazon Rekognition, see Detecting objects and concepts in the service documentation.
For tasks that require the standard labels in Amazon Rekognition, testing this service is worthwhile. If Amazon Rekognition can meet your requirements, the model selection, training, and maintenance are abstracted. It provides a pretrained service for inference, and AWS handles maintenance of the service. Obtaining predictions from Amazon Rekognition is straight-forward.
The following are the advantages of Amazon Rekognition:
-
Immediately available and scalable
-
No training or configuration required
-
Supports multi-label classification
The following are the disadvantages of Amazon Rekognition:
-
Fixed set of predicted classes
-
Inference units offer chunks of capacity, and the smallest unit might be costly for small throughput
For more information, see the following:
-
Getting started with Amazon Rekognition in the Amazon Rekognition Developer Guide
-
DetectLabels in the Amazon Rekognition API Reference