Abstract: A check processing equipment having an optical amount character recognizing function for a proof machine used at a bank. A received amount can be corrected on an image screen, and it enables the efficient processing for certifying operation. A document capturing unit converts the data on the document into an image data and transfers the image data to a frame memory of recognizing unit. A recognizing unit compresses the image data through a recognizing/determining section and also carries out CAR with the attribute on an E13B code line data. A work station stores the compressed image data, the code line data, and the CAR result in a hard disk and it also displays an extended image and implements the proof of a whole batch through key-in operation for handling unrecognizable data and exceptional error data. An encoder unit performs encode printing and endorse printing on proved documents while verifying the code lines thereof.

Abstract: A plurality of recognition candidates are obtained through continuing a segmentation to a numeral recognition realized by segments, and an optimum one is selected from among the candidates by means of connection information, thereby preventing continuous numerals from being recognized erroneously. A method of correctly recognizing the handwritten characters, especially numerals written in checks includes the steps of: performing a line drawing analysis is of the character set to be recognized in order to extract a characteristic of the arrangement of blocks, proposing a hypothesis about the monetary unit type, and finally verifying the above hypothesis to display the result of the character recognition. Further, this method can surely distinguish handwritten characters from a bar which is either printed for entry of characters or handwritten, and eliminate the bar from the object of character recognition to extract the characters alone.

Abstract: A letter pattern recognition method preprocesses each character pattern to effect characteristic extraction based on elements such as a loop, a straight line and an arc and is adapted to compare the character pattern with a preregistered classification code after the normalization thereof, the character pattern being subjected to balance and stroke checks if necessary as a result of the comparison. Notwithstanding this, if the character is not defined, the character pattern is identified by means of detail classifications such as those of a divergent mode of the arc, and a length ratio of the elements and the like.