Human review in Textract

Short answer No

Amazon Textract is an automated text recognition service that uses machine learning to extract text and data from scanned documents. When you use the Human Review feature with Amazon Augmented AI (A2I) and Amazon SageMaker, you are adding a layer of human review to the process of text recognition. This means that a human worker will review the output of the Textract model and make any necessary corrections.

The Textract model does not directly learn from the human review process. However, you can use the output of the human review process to train a separate machine learning model that can learn from the corrected data and improve its performance on similar documents. This is known as supervised learning, where you provide labeled training data to the model so that it can learn to make predictions on new, unseen data.

To do this, you would need to export the output of the human review process, which includes the original text that was recognized by Textract as well as any corrections made by the human worker, and use this data to train a new machine learning model. You can then use this model to perform text recognition on new documents, and it should be able to make more accurate predictions as a result of having learned from the corrected data.

It's important to note that the effectiveness of this approach will depend on the quality and quantity of the training data that you use to train the new model. The more data you have and the more diverse and representative it is, the better the model will be able to generalize to new, unseen data.

@seekrsi