17-03-2016, 04:09 PM
19-03-2016, 03:13 PM
character segmentation code in java
Plot the histogram (gradient value vs number of black pixel) and find the segmentation point for splitting the cursive character.
1.0 For each pixel column in the text line
1.1 Count the number of black pixels
End For
2.0 Generate histogram representing the number of black pixels in each column
3.0 Find the maximum value in the histogram
4.0 Set the threshold using the maximum value in the histogram
5.0 Assign positive and negative signs to the histogram gradient values.
6.0 For each group of consecutive negative sign values in the histogram
6.1 Count the negative sign values
6.2 If the number of negative values is less than the threshold
6.2.1 Convert all the negative sign values to positive
End If
End For
7.0 Set the text line height as the sliding window width
8.0 For each histogram value
8.1 If negative sign gradients meets positive sign gradients
8.1.1 Set pixel count as character segmentation point
End If
End For
9.0 For each character segmentation point
9.1 Find next segmentation point
9.2 If length between the segmentation points is less than the sliding window width
9.2.1 Segment based on the segmentation points
End If
9.3 Else
9.3.1 Segment based on the sliding window width
End Else
End For
Plot the histogram (gradient value vs number of black pixel) and find the segmentation point for splitting the cursive character.
1.0 For each pixel column in the text line
1.1 Count the number of black pixels
End For
2.0 Generate histogram representing the number of black pixels in each column
3.0 Find the maximum value in the histogram
4.0 Set the threshold using the maximum value in the histogram
5.0 Assign positive and negative signs to the histogram gradient values.
6.0 For each group of consecutive negative sign values in the histogram
6.1 Count the negative sign values
6.2 If the number of negative values is less than the threshold
6.2.1 Convert all the negative sign values to positive
End If
End For
7.0 Set the text line height as the sliding window width
8.0 For each histogram value
8.1 If negative sign gradients meets positive sign gradients
8.1.1 Set pixel count as character segmentation point
End If
End For
9.0 For each character segmentation point
9.1 Find next segmentation point
9.2 If length between the segmentation points is less than the sliding window width
9.2.1 Segment based on the segmentation points
End If
9.3 Else
9.3.1 Segment based on the sliding window width
End Else
End For