Prime Vision has developed a machine recognition system, Context Reco, to provide a solution for reading both hand and machine written text. Context Reco is also compatible with both ParcelMatch and MailMatch open architecture platforms.
The system acquires around 50 per cent of the information from the character itself, but the other 50 per cent comes from the surrounding text. This is information such as position or size in relation to the line.
Context Reco has been based on the way humans make sense of text by not performing these tasks sequentially, but co-ordinated. When tested with different language sets it was found that 15 per cent less video coding was required.
