Hi All,
sorry for the English. I can read German, but fear if I tried to write it, it would be unreadable and an insult to the language.
I recently created an OCR utility for phone numbers. It is not intended as a full dialer, but instead as a utility that can be used as a plugin by a dialer. Specifically, I wrote it with VoIPdial in my mind, which is a great utility that can be used on any text that you can select with the mouse.
But sometimes you cannot select text, because either the text is not select-capable or because the text is active and reacts to the selection. Or because you are too lazy to be bothered (I know I am
)
So PhoneOCR can be used anywhere the text is visible. Just 'hover' the mouse over the phone number when you activate the program. It will next OCR the phone number, optionally present a validation box, and return the phone number by standard-out or by text file, to be used by the dialer.
=>
For now, I have it in a batch script (with Windows shortcut key) that next calls VoIPdial, but VoIPdial author has indicated he may integrate the program better, once he has dealt with the SID problem, which is a higher priority.
----------
Usage is very simple: Restore the content of the file in a directory, probably a plugin directory from your dialer program. From a command line, or a script, or whatever, call program phoneocr.exe with or without options. Read the PhoneOCR.cmd for hints on a batchfile wrapper
Use phoneocr -h for more help.
Use phoneocr -d to collect debug information and look at intermediate results if you do not understand why it is mis-scanning a number
The program does need .net (v2 I think) and GOCR, the latter is included. It is really GOCR that does the OCR, PhoneOCR merely tries to apply it to phone numbers.
----------
A warning though: OCR is not without error. I have done my best to filter out all errors by OCR'ing the text not once, but three times, with different parameters, and sort through the resulting mess. Run with -d option if you want to see the sometimes messed up intermediate results. (for example, if it sees a "J" in one scan, a "1" in another, and a "_" in the 3rd scan, then I know it is supposed to be a "7")
But it does its work, as long as the text is not too small or times-roman or bold or italic. And if it is small-bold-italic-times-roman, then it never works...
----------
Future of this utility: Well, there may not be one. But if popular demand request so, it is possible to extent sensitivity by creating a specific character databases. Also, if you want to have access to the source (visual basic) you can request this, but be warned, with all the special cases and workarounds, the code has become a bit of a mess. A full rewrite in C++ is what I would like to do, but I already spend too much time on this.
sorry for the English. I can read German, but fear if I tried to write it, it would be unreadable and an insult to the language.
I recently created an OCR utility for phone numbers. It is not intended as a full dialer, but instead as a utility that can be used as a plugin by a dialer. Specifically, I wrote it with VoIPdial in my mind, which is a great utility that can be used on any text that you can select with the mouse.
But sometimes you cannot select text, because either the text is not select-capable or because the text is active and reacts to the selection. Or because you are too lazy to be bothered (I know I am
)So PhoneOCR can be used anywhere the text is visible. Just 'hover' the mouse over the phone number when you activate the program. It will next OCR the phone number, optionally present a validation box, and return the phone number by standard-out or by text file, to be used by the dialer.
For now, I have it in a batch script (with Windows shortcut key) that next calls VoIPdial, but VoIPdial author has indicated he may integrate the program better, once he has dealt with the SID problem, which is a higher priority.
----------
Usage is very simple: Restore the content of the file in a directory, probably a plugin directory from your dialer program. From a command line, or a script, or whatever, call program phoneocr.exe with or without options. Read the PhoneOCR.cmd for hints on a batchfile wrapper
Use phoneocr -h for more help.
Use phoneocr -d to collect debug information and look at intermediate results if you do not understand why it is mis-scanning a number
The program does need .net (v2 I think) and GOCR, the latter is included. It is really GOCR that does the OCR, PhoneOCR merely tries to apply it to phone numbers.
----------
A warning though: OCR is not without error. I have done my best to filter out all errors by OCR'ing the text not once, but three times, with different parameters, and sort through the resulting mess. Run with -d option if you want to see the sometimes messed up intermediate results. (for example, if it sees a "J" in one scan, a "1" in another, and a "_" in the 3rd scan, then I know it is supposed to be a "7"
)But it does its work, as long as the text is not too small or times-roman or bold or italic. And if it is small-bold-italic-times-roman, then it never works...
----------
Future of this utility: Well, there may not be one. But if popular demand request so, it is possible to extent sensitivity by creating a specific character databases. Also, if you want to have access to the source (visual basic) you can request this, but be warned, with all the special cases and workarounds, the code has become a bit of a mess. A full rewrite in C++ is what I would like to do, but I already spend too much time on this.
