Knowledge check

Answer the following questions to check your learning.

In this module, our image data was referenced using a datastore that attached to an existing ...

Azure Blob Storage in an Azure Storage account

Azure Synapse Analytics data pipeline

File Share in an Azure Storage Account.

ML-assisted labeling was not required in this module, how might it be beneficial to a data labeling project?

ML-assisted labeling creates a model that can pre-label data in new image samples that may be reviewed later by team members for accuracy.

ML-assisted labeling can determine objects in an image using an existing model that detects a variety of common objects.

Allows you to submit unlabeled to be labeled by a third party service that employs human labelers.

The reason we labeled 10 images in our dataset was to satisfy the following requirement:

10 is the total number of images that make up our dataset

10 is the maximum number of images that can be used to train an object detection model.

Azure ML Studio Object Detection Experiments require a minimum of 10 labeled samples in order to successfully train our object detection model

You must answer all questions before checking your work.

Feedback