AlfredOCR - Optical Character Recognition

[TomBenz]

@zeitlings This is an amazing workflow and super fast. Many thanks for it.

 

Alfred OCR+ -- when extracting data from images such as https://exceljet.net/formulas/all-values-in-a-range-are-at-least here it gives column data one after the one or sometimes skips zero or single-letter Alphabet. Is it possible to check and see if anything can be done? For now, I have to select each region or column seperately and then manually correct.

[zeitlings]

Hey @TomBenz

 

What's your desired result? A plain text representation that mimics the table?

Currently the coordinates of the extracted text are not preserved, and I don't see an easy way to achieve that.

If some single characters are not recognized, that's something that Apple will have to improve in their Vision model, there's really nothing we can do about it. That empty cells are skipped is expected behavior, since the program only targets recognized text and knows nothing about the tabular format.

[TomBenz]

2 hours ago, zeitlings said:

Hey @TomBenz

 

What's your desired result? A plain text representation that mimics the table?

Currently the coordinates of the extracted text are not preserved, and I don't see an easy way to achieve that.

If some single characters are not recognized, that's something that Apple will have to improve in their Vision model, there's really nothing we can do about it. That empty cells are skipped is expected behavior, since the program only targets recognized text and knows nothing about the tabular format.

@zeitlings Thanks for your reply. I am not sure if it is feasible but I would like to paste the result in seperate column of excel. Excel has a feature that generate data from picture. It gives data in better format  keeping columns etc. but it is very slow. So if there is a way to preserve coordinates, it would be great. Empty cells are fine and skipped -- this is fine and not a problem. But it also skips single character like A or B in cell. Also it remove some consecutive duplicates. 

[zeitlings]

Hey @TomBenz

For now, I won't try to implement this particular feature. As you can imagine, it's a lot more complex than doing simple OCR. However, I will keep it in mind and revisit it if I come across a feasible solution or if a similar need arises in another context. At least Excel is already doing what you need, albeit very slowly 😄

[TomBenz]

4 hours ago, zeitlings said:

Hey @TomBenz

For now, I won't try to implement this particular feature. As you can imagine, it's a lot more complex than doing simple OCR. However, I will keep it in mind and revisit it if I come across a feasible solution or if a similar need arises in another context. At least Excel is already doing what you need, albeit very slowly 😄

thank you.