Apparatus and method for separation of optical character recognition data

The method of separating character data of a first and second adjacent character in a data field having vertical columns and horizontal rows of black and white bit data representative of the image of characters in a horizontal row of type, comprising the steps of:

• a. evaluating vertical columns of data to determine whether a column having all white bits exists at the area of the character data most probably including the midpoint of the two characters and if so separating the data on either side of said column;
• b. if a vertical white column does not exist, next evaluating the character data to determine whether a continuous path of white bits exists having minimum contiguity of a first plurality of bits vertically and a second plurality of bits horizontally;
• c. if a continuous white column exists, separating the data in each data row on either side of the path by determining which black bits within a row are contiguous with black bits in the same row, one row above or one row below in the adjacent column and which black bits are contiguous with black bits just above or below with the same data column and storing all of said bits with data from the first character and storing all remaining bits with data from the second character;
• d. if neither a vertical all-white column nor a continuous white bit path having the required contiguity pattern exists, next counting the number of black bits in each of a selected group of adjacent vertical columns selected to be in the vicinity of the character data which most probably includes the mid-point of said adjacent character;
• e. adding to the count of the extreme column in the group of adjacent vertical columns a preselected number;
• f. determining among the black bits count for each column including the added counts which column has the lowest number of bits;
• g. determining which of the two columns on either side of the least count column has the lowest count, and;
• h. grouping data which is located to the side of the lowest count column on which the next lowest count column is located into data from the first character and data on the opposite side into data from the second character.

An apparatus for evaluating a digital data field, which field representative of a series of alphanumeric characters, some of which may be underlined, for the purpose of detecting and removing said underlining, said data field having a first plurality of columns, each column in turn having a second plurality of data bit positions, each of which positions may contain data bits representative of the presence of character or underline information, referred to as "black bits", or may contain data bits representative of the absence of character data, referred to as "white bits", comprising:

• a. digital memory means for at least temporarily storing digital data;
• b. first memory scanning means responsive to said digital memory means for sequentially scanning said first digital memory means and detecting the presence of all white bits in the same position in a pre-established number of contiguous columns, and having an output signal which indicates that such a condition has occurred;
• c. second memory scanning means, responsive to said digital memory means for sequentially scanning said digital memory means for detecting the presence of all black bits in the same position in the pre-established number of the same said contiguous columns, and having an output signal which indicates that such a condition has occurred;
• d. sequence detection means having inputs responsive to the outputs of said first memory scanning means and said second memory scanning means and itself having an output which indicates the detection of a sequence of output signals from the first and second memory scanning means in which an output signal from the first memory scanning means is followed by an output signal from the second memory scanning means which exists for no more than a pre-determined number of contiguous data bit positions and which is in turn followed by another output signal from said first memory scanning means;
• e. memory position tabulation means responsive to the output of the sequence detection means for identifying the memory positions in which all black bits which have caused an output to occur from the sequence detection means are located, and;
• f. logic means responsive to the memory position tabulation means for selectively removing from said data field all black bits which have caused an output to occur from the sequence detection means.