Create a new OCV tool based on gray value projections.

create_ocv_proj creates a new OCV tool. This tool will be used to train good-patterns for the optical character verification. The training is done using the operator traind_ocv_proj. Thus traind_ocv_proj is normally called after create_ocv_proj.

The pattern comparison is based on the gray projections: For every traing pattern the horizontal and vertical gray projections are calculated by summing up the gray values along the rows and columns inside the region of the pattern. This operation is applied to the training patterns and the test patterns. For the training patterns the result is stored inside the OCV tool to save runtime while comparing patterns. The OCV is done by comparing the corresponding projections. The Quality is the similarity of the projections.

Input for create_ocv_proj are the names of the patterns (PatternNames) which have to be trained. The number and the names can be chosen arbitrary. In most case only one pattern will be trained, thus only one name has to be specified. The names will be used when doing the OCV (do_ocv_simple). It is possible to specify more names than actually used. These might be trained later.

To close the OCV tool, i.e. to free the memory, the operator close_ocv is called.

create_ocv_proj("A",&ocv_handle);
draw_region(&ROI,window_handle);
reduce_domain(Image,ROI,&Sample);
traind_ocv_proj(Sample,ocv_handle,"A","single");

create_ocv_proj returns 2 (H_MSG_TRUE), if the parameters are correct. Otherwise, an exception handling is raised.

create_ocv_proj is processed completely exclusively without parallelization.

traind_ocv_proj, write_ocv, close_ocv

read_ocv

create_ocr_class_box

Optical character verification