This part involves measuring those relevant features of
the cleaned-up input character of the upper line, and
province-name block of the lower one. After feature
extraction, the pattern of each character and the pattern of
province-name block are represented by the set of extracted
features. For our CLP, the type of interesting feature is
black-color intensity. Each character pattern and a single
province-name block pattern will be divided into 5 x 5
blocks as shown in figure 5. For each small block, blackcolor
intensity will be counted in terms of the numbers of
pixel. This can be done as in equation (3).