By using AWS re:Post, you agree to the AWS re:Post Terms of Use
AWS re:Postre:Post

How to Boost Performance of Amazon Textract Table Extraction

1

Hi I am using Textract on images of tables and noticed quite poor performance on denser tables. Are there any parameters I can tweak to boost this performance in the example below? I'm surprised that the AI cannot tell how gridlines divide cells.

I'm using Python API with the below:

extractor = Textractor(region_name="us-east-1") # Initialize textractor client, modify region if required

document = extractor.analyze_document(
    file_source=image,
    features=[TextractFeatures.TABLES],
    save_image=True #Keep this true to see bounding boxes for testing
)

Image that Textract fails at

Topics
Machine Learning & AI
Tags
Amazon Textract
Language
English
dmilo123
asked 9 months ago200 views
2 Answers
0

Thank you for bringing this issue to our notice. With machine learning models, we cannot guarantee 100% accuracy. We are continuously improving the accuracy of our models in response to our customer feedback. Please share at least 10 documents that are representative of your production traffic to help us better diagnose the issue and also refer to our best practices to optimize how you use Textract.

AWS
AWS-User-3345173
answered 9 months ago
  • dmilo123
    9 months ago

    I posted 10 examples in this thread

0

Sure, here are 10 examples.

Enter image description here Enter image description here Enter image description here Enter image description here Enter image description here Enter image description here Enter image description here Enter image description here Enter image description here Enter image description here

dmilo123
answered 9 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions

Relevant content