Work to do:
Extract information from tables in invoices.
The table can expand/grow vertically depending on how much the customer has bought. more products on the invoice means more lines and the table expands.
Less means a shorter table.
In the attached example you can see a table with two lines. The same company can send tables with 20 lines
Some other companies do send tables with no lines, or less lines.
The vertical size of the invoice affects where to find the text to extract.
This also affects the bboxes below for other fields.
This i opencv, tesseract, parsing PDF through OCR job.
Please tell me how you would solve the invoices and tell me how to do this.
we give you helper_headers and helper text, then you detect the rows. You detect the rows by analysing small snippets of the width of the paper sheet and compare how much black and white you have. the lines that create a new row have more black on the total space then the ones that only have some text description.
So your job is to detect where it is a new row.
We give you the columns
you find the rows.
There are some more tricks that can be applied too , but lets start with the above.
so the job is very simple if you are a vision EXPERT (a real one...). So I expect you to have a Phd in it or something similar so you know your stuff properly.
2. Interviewing and experience
We must interview you and we need to see examples of your code or if you have done the job you can run through the examples and show you can detect the table.
Please refrain from saying "I will do it in best way possible", "I am an expert", "I have 500 developers and we have high ranking" anyone knows you cannot have grade 5 as an average value so it is worth nothing. No one get grade 5 as an average on 5+ or even 200 jobs. And yes, I have been through a few tricks here too so I know how the system works.
Read the advert and tell me you what experience you have in the area.
If you send a "bla bla" response or automated with the header included I am just removing you.
Show quality, show how you separate from the others by telling experience in detail and that you have read the advert (no, saying "I have read and understood everything" shows you cannot read and are useless)
Please note, you do the job with commited and acxcepted project and not as a guest for a few days and then you abandon the project.
Price is a placeholder. It means you can bid higher or lower
Please note. If you lie about your nationality, please expect to be disqualified immeditely. That it is directly aimed at Chinese who pretend having other nationalities. We do recognize your accent.