Alteryx intelligent Suite - Computer vision

[sriniprad08]

Hi Team,

Hope you are well. I am currently exploring on extracting information from the invoices using the Alteryx intelligent suite - computer vision. But i am unable to see any workflow developed or any proper training material to execute this.

Please let me know if  any one has experience in developing this and can share the workflow or relevant materials would be of great help,

Thanks in advance,

Regards,

Sri

[mceleavey]

Hi @sriniprad08 ,

Absolutely. Unfortunately, this will convert the invoice into text, and so you will require different parsing for each template format.

Simply load int he image, I assume pdf, using the image input tool, then use image to text. This will convert each page of document into a single cell of text, which you then parse out using the regex/text to columns tools.

If you have an example invoice I can quickly build the basics for you.

M.

[sriniprad08]

Hi @mceleavey ,

Thank you so much for the kind reply and offer to help. Please find attached the sample invoices.

One question please so if we have 1000 invoices with different format and if we would like to extract only the date, invoice number and amount. Do we need to create workflow separately for all of them? or is there way to extract using the alteryx intelligent suite?

Thank you again,

Sri

[mceleavey]

Hi @sriniprad08 ,

You can use two methods. The first is a bit hit and miss, which is loading in your image, then feeding this into the Image Template tool:

This tries to detect tabular or boxes of information:

The second is to use the Image Template tool without a stream coming in. Manually select the invoice and manually select the information in the image:

You can embed this in a batch macro to overwrite the manual file selection and build a template per supplier.

However, there is no way of automatically pulling the information from thousands of different templates.

Hope this helps,

M.

[sriniprad08]

Hi @mceleavey ,

Thank you very much for the help. I understand. Is it possible for you to share the workflow please?

Cheers,

Sri

[cgoodman3]

Here’s a video I did when AIS was new, there’s been lots of cool updates such as table detection when you read in pdfs (what @mceleavey showed).

https://m.youtube.com/watch?v=63Dvm3M1Uak

[CristonS]

hey everyone, FYI there is a Learning Path for Intelligence Suite, including the Computer Vision tools. Also check out these blogs: Computer Vision Within Reach and Image Recognition: Classification Models Made Simple. Thanks for the feedback about wanting more learning resources!