Hi,
We are evaluating Abbyy's OCR after having no success with Tesseract scanning dot matrix codes.
We print unique codes that go out on pack to a consumer. These codes can then be entered on our site by the consumer to redeem against prizes etc.
We want to allow the user to use their camera to scan the codes to save them having to type them all in manually. When you have 10's of codes this can be time-consuming.
The codes are printed in dot-matrix, using a known length and character set.
Our first test with the cloud OCR has turned out disappointing, to say the least. See the attached image for the result.
We are using processTextField and this is the param URL we used:
https://cloud.ocrsdk.com/processTextField?letterSet=CEFHKPTWXY®Exp=(C|E|F|H|K|P|T|W|X|Y){14}&textType=matrix&oneTextLine=true&oneWordPerTextLine=true
Is there anything wrong we are doing there that is giving us a poor result or anything we can do to improve the result?
Also, has anyone had any experience using Abbyy OCR successfully scanning codes?
コメント
1件のコメント
Hi Scott!
At first, we would like to notice that curly braces in regular expressions are not supported by Cloud OCR SDK. Please see the full specification of used signs on the following page: Regular Expressions. You can replace the braces by repeating the particular construction the required number of times.
Also please pay your attention to the fact that regular expressions do not strictly limit the way how the result looks. During recognition, all hypotheses of a word recognition are checked against the specified regular expression. If a given recognition variant conforms to the expression, it has a higher probability of being selected as final recognition output. But if there is no variant that matches the regular expression, the result will not conform to the expression.
If it is possible, please send more sample images to cloudocrsdk@abbyy.com. It will make us able to assist you better.
サインインしてコメントを残してください。