ABBYY OCR SDK

ABBYY SDK 'Expert' We are looking for a talented person who knows ABBYY SDK OCR software. We need a service where a user 'selects a rectangle' on a picture. (i'm thinking of something with javascript) After selecting this rectangle the software gets a list of hunderds of similar images ( remote png files ) to be processed with ABBYY Field level recognition (using the previously selected coordinates). It need to return the recognized text and a confidence level. These things are relativetely easy with the well documented class at [login to view URL] We prefer PHP or a commandline script. (ofcourse the 'select rectangle' is an online page, but the processing can be in the background) If you have ideas let us know. Please give me a reply, quote and if you have experience with the sdk.