A Neural Approach to Cursive Handwritten Character Recognition Using Features Extracted from Binarization Technique - Complex System Modelling and Control Through Intelligent Soft Computations - page 742

Fig. 10 Resized handwritten
character image

5.5 Size Normalization

Since handwriting depends on the writer and even a single person cannot always
write the same character in exactly the same way. The shape of the handwritten
characters may be different under different conditions such as speed of writing,
position of writer (sitting or standing) while writing and quality of pen and/or paper
etc. The same character image (say

) written by different writers or by the same
writer may be of different sizes. Also, there is a size difference among all the 26
characters present in the English Language. Some characters are of different height
and some are of different widths. Cropped character images obtained by removing
the extra space in the image around the character are also of different sizes.
There is a need to put all the handwritten character images in a uniform size i.e.
the characters should be in normal form. To make all the character image samples in
the normal form, all the character images are reconstructed in the size of 15

' c '

×

12

pixels by using method of MATLAB which employs nearest neigh-

borhood interpolation technique. The character image after resizing is shown in
Fig. 10 .

' imresize '

6 Feature Extraction

Feature extraction is a process of studying and deriving useful information from the

filtered input patterns. The derived information can be general features, which are
evaluated to ease further processing. The selection of features is very important