Computer Vision - Calculating Extended Cost

[oracleoftemple]

See below please.  This is just a snapshot of what I'm dealing with.  Not exactly high-res images.  The three columns on the right are what's most important.  Qty, Rate, and Total.  Total is just the extended cost of Qty x Rate.  I'm not only interested in extended cost.  The quantities and rates are also important.  My question is this - if OCR has trouble determining exactly what numbers are in any of those three columns, can the formula Qty x Rate = Total be used to fill in the gaps for numbers that are unreadable?  For example, if OCR determines that the Total is $734.91 and the rate is $28.05/hr but it can't quite make out what the quantity is, does Alteryx have a way to compute that field by dividing the total by the rate?  I know I could just import the data and then attempt to fill in the gaps, but I was hoping to make this as automated as possible.  There are lots of files

[Matt_D]

Hi @oracleoftemple how do you know that any of the fields have been read correctly? That is where your challenge lies with Computer Vision or any OCR capability. The image has to be a certain quality and alignment to be translated from image to data accurately.

You can build out the calculations very easily after the OCR process has ran and try and use them to validate what was read from the image but that's not the say the other values have been read correctly either.

If you know you can read 2 of the values more consistently than the 3, and can derive the 3rd based on the 2 why wouldn't you do this anyway? It would eliminate some risk.

I know this might not help but I don't think there is a robust solution for you here aside from making sure the images you're reading are perfect.