Text and Image analysis

George Nicholson 6

I would like to use azure to analysis multiple bank statements and classify transactions, how do I go about doing this?

3 answers

Ramr-msft 17,611 Reputation points

2020-11-12T18:43:40.703+00:00

@George Nicholson Thanks for the question, Please follow the OCR and Form Recognizer seems promising for your application. Can you please add more details about the sample inputs and labels.

Please follow the kiosk samples.

Kiosk sample on GitHub: https://github.com/Microsoft/Cognitive-Samples-IntelligentKiosk/
Please sign in to rate this answer.
Ramr-msft 17,611 Reputation points

2020-11-30T06:08:10.597+00:00

@George Nicholson Just checking to see any update on this. If the suggested response helped your issue, do click on "Mark as Answer" and "Up-Vote" for the answer that helped you for benefit of the community.
Sign in to comment
George Nicholson 6 Reputation points

2020-12-04T14:27:54.59+00:00

Hi
They will be a mix of Jpegs and PDFs and other image types.
Please sign in to rate this answer.
Ramr-msft 17,611 Reputation points

2020-12-14T15:22:39.57+00:00

@George Nicholson There’s a solution accelerator for Knowledge Mining that could fit your case: https://github.com/microsoft/Accelerator-AzureML_CognitiveSearch
It combines reading text from documents using Azure Search’s OCR capabilities (as suggested below) + training and deploying a Natural Language Processing model using Azure Machine Learning.

In the example the model is doing Named Entity Recognition, not classification, but you could replace it by a classification model. Reference materials I can think of:

Automatically train a classification model using AutoML in AML: https://github.com/Azure/MachineLearningNotebooks/blob/master/how-to-use-azureml/automated-machine-learning/classification-text-dnn/auto-ml-classification-text-dnn.ipynb

Training a custom classification model: check NLP recipes https://github.com/microsoft/nlp-recipes/tree/master/examples/text_classification
Sign in to comment
George Nicholson 6 Reputation points

2020-12-04T14:33:44.22+00:00

I also dont really get how to start building this with the OCR and Form recogniser?
Please sign in to rate this answer.
Ramr-msft 17,611 Reputation points

2020-12-14T15:19:18.563+00:00

@George Nicholson Thanks, If you are extracting only text, tables and selection marks from documents you should use layout, if you also need to extract key value pairs you can train a custom model or use a pre-built (Invoice, Receipts, Business Cards). Layout results (text, tables and selection marks) are included in all the Analyze outputs (custom and pre-built) in the readResults (text) and pageResults (tables) of the JSON output.

• Layout – extract text, tables selection marks no training required
• Pre-built – Invoice, Receipts, Business Cards – extract values of interest from these type of documents
• Custom – Extract key value pairs trained on your own documents
All of the above will also include the text, tables and selection marks in the results.
Sign in to comment