Method of separating text and graphs in digital image data
View Patent ↗A method of separating text and graphs in digital image data is used to separate digital image data into graph data and text data. According to a default separating parameter, the digital image data are partitioned into several operating matrices. One then sets a default crossing reference level for comparison analyses to compare adjacent pixel units in each operating matrix. When the digital characteristic value of a pixel unit is greater than the default crossing reference level and the digital characteristic value of its one adjacent pixel unit are smaller the default crossing reference level, the operating matrix is given a weight. Afterwards, all the weights are added up to give a weighted statistical value, which eventually is used to separate the graph data from the text data.
1. A method of separating graphs and text in digital image data for separating a plurality of graph data from a plurality of text data contained in a digital image, the method comprising the steps of:
reading the digital image data;
partitioning the digital image data into a plurality of operating matrices according to a default separating parameter, each of the operating matrices consisting of a plurality of pixel units each of which contains a digital characteristic value;
setting a default crossing reference level (CRL); and
for each of the operating matrices performing the following steps:
performing a comparison analysis, which compares each of a plurality of pairs of adjacent pixel units in the operating matrix according to the default CRL and the operating matrix is given a weight when the digital characteristic value of the a first one of the adjacent pixel units is greater than the default CRL and the digital characteristic value of a second one of the adjacent pixel units is smaller than the CRL, and the operating matrix is given the weight when the digital characteristic value of the first adjacent pixel unit is smaller than the CRL and the digital characteristic value of the second adjacent pixel unit is greater than the CRL,
adding up the weights of the operating matrices to obtain a weighted statistical value, and
defining the operating matrix as text data when the weighted statistical value is smaller than a separating reference level (SRL) and as graph data when the weighted statistical value is greater than the SRL.
2. The method of claim 1 , wherein the step of performing a comparison analysis is done for two adjacent pixel units in the X direction.
3. The method of claim 1 , wherein the step of performing a comparison analysis is done for two adjacent pixel units in the Y direction.
4. The method of claim 1 , wherein the operating matrix consists of N×N pixel units, each of which contains a set of coordinates.
5. The method of claim 4 , wherein the coordinates consist of an X-axis coordinate value and a Y-axis coordinate value.
6. The method of claim 1 , wherein the operating matrix contains an operation attribute.
7. The method of claim 6 , wherein the operation attribute is set as false when the operating matrix contains text data.
8. The method of claim 6 , wherein the operation attribute is set as true when the operating matrix contains graph data.