Need a PHP/CI or SQL solution for a complex issue (only apply if you have a solution) -- 2
Budget: $30 – $250 USD
Please ONLY bid on this project if you have a clear and tested solution to the following issue:
We are submitting pdf documents into Tesseract for OCR. This produces a string output of the raw text obtained by the OCR process.
Let us say that for our purposes, we want to identify if any of the Unique Customer ID numbers in our DB are contained in this output text. We have all of this working - no issues.
Here's the problem. OCR is not 100% accurate. So a customer ID in our DB might be CID000115, so it is checking the output text for this number, however maybe the OCR returned CID0ll5 in the output string. Meaning - the OCR process detected "l's" instead of "1's".
So here is the solution we need to implement. We are creating a table with all of the common mistakes by the OCR. So in this case, we will have one table record specifying that 1 could be interpreted as l. And many other common mistakes the OCR might make we will be adding as we identify them.
So here is the logic we need to accomplish:
Search OCR_Table.OutputString for all Customer ID's in our customer table but allow for any substitutions from our common ocr errors table.
So when searching the text string, it should find CID000115, or CID000ll5 (because in our table we know that 1's can be mistaken for l's. There will also be many other possible characters in our table that are common errors, so for example if "0" can be mistaken for "O" and this is in our errors table, then if the output string includes CIDOOO115, it would also find this as true.
We need to be able to search the output string for EVERY SINGLE CID in our database WITH ALL POSSIBLE VARIATIONS from the ocr error variations table reference.
If you have a solution to this in either PHP/CI or SQL, we will award you the project. Feel free to ask any questions.
Thank you.
*********NOTE: We prefer that the entire solution be done in SQL as to not add too much load to the web server*******
We are submitting pdf documents into Tesseract for OCR. This produces a string output of the raw text obtained by the OCR process.
Let us say that for our purposes, we want to identify if any of the Unique Customer ID numbers in our DB are contained in this output text. We have all of this working - no issues.
Here's the problem. OCR is not 100% accurate. So a customer ID in our DB might be CID000115, so it is checking the output text for this number, however maybe the OCR returned CID0ll5 in the output string. Meaning - the OCR process detected "l's" instead of "1's".
So here is the solution we need to implement. We are creating a table with all of the common mistakes by the OCR. So in this case, we will have one table record specifying that 1 could be interpreted as l. And many other common mistakes the OCR might make we will be adding as we identify them.
So here is the logic we need to accomplish:
Search OCR_Table.OutputString for all Customer ID's in our customer table but allow for any substitutions from our common ocr errors table.
So when searching the text string, it should find CID000115, or CID000ll5 (because in our table we know that 1's can be mistaken for l's. There will also be many other possible characters in our table that are common errors, so for example if "0" can be mistaken for "O" and this is in our errors table, then if the output string includes CIDOOO115, it would also find this as true.
We need to be able to search the output string for EVERY SINGLE CID in our database WITH ALL POSSIBLE VARIATIONS from the ocr error variations table reference.
If you have a solution to this in either PHP/CI or SQL, we will award you the project. Feel free to ask any questions.
Thank you.
*********NOTE: We prefer that the entire solution be done in SQL as to not add too much load to the web server*******