Post-recognition segmentation for pattern-recognition machines

An optical scanner converts input images into video snapshots which are recognized at fixed intervals, independently of the presence of patterns. Sense generators identify the presence of pattern left and right sides in the snapshots, independently of the recognition unit. A direction detector determines whether the scanner is moving from left to right or vice versa. A composition unit converts the fixed-interval recognitions into a single output code for each input pattern, eliminating duplicate and incorrect data.