You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I dont really know how this magic works, but as i understand, there is a possibility array for each symbol it found. I often get "B" instead of "3".
So i propose to make an ability to choose between character sets, so when it thinks that its 95% "B" and 90% "3" its gonna choose "3" instead of "B"
Maybe you will add this in TODO list
The text was updated successfully, but these errors were encountered:
If you're only limiting the character set, you can use the characterBlackList and characterWhiteList properties, which work as advertised. For example:
let ocr = SwiftOCR()
ocr.characterWhiteList = "1234567890"
ocr.recognize(ocrImage) { string in
print(string)
}
If you want to add characters to the list (e.g. lowercase letters, as in #122, or punctuation, as in my case), then it seems you have to train a new network. That part doesn't seem to be working as well for me.
I dont really know how this magic works, but as i understand, there is a possibility array for each symbol it found. I often get "B" instead of "3".
So i propose to make an ability to choose between character sets, so when it thinks that its 95% "B" and 90% "3" its gonna choose "3" instead of "B"
Maybe you will add this in TODO list
The text was updated successfully, but these errors were encountered: