Feature Extraction Methods Algorithms1. Intersection PointsPrinciple:• The number and position of intersections between the character and straight lines have been retained as features.• 1st horizontal (1/3) height.;• 2nd horizontal (2/3) height. ;• The vertical straight line passes through the center of gravity of the characterAlgorithm:INPUT: image of the characterOUTPUT: number and position of intersections between the character and the straight lines begin Step 1: transform the input image into a thinned image Step 2: Draw two horizontal straight lines in the first third and second third of the character's height Step 3: The vertical straight line passes through the character's center of gravity Step 4: Find the X position of the intersection point of the horizontal line at 1 /3 of the character's height with the character. Step 5: Find the X position of the intersection point of the horizontal line at 2/3 of the character's height with the character. Step 6: Find the Y position of the intersection point of the vertical line at the character's center of gravity with the character.Step 7: Store the above features in a file. end • No. of horizontal intersection points 2 • No. of vertical intersection points 1 • Total no. of intersection points 32. Characteristics of the zone centroid Principle: • The main advantage of this approach comes from its robustness to small variations, ease of implementation and good recognition rate. • The zone-based feature extraction method provides good results even when some pre-processing steps such as filtering, smoothing and skew removal are not considered. Algorithm 1: Image centroid and ICZ (Distance Metric Feature Extrac) based on zones...... center of sheet ...... horizontal profile =( left side view )/( right side view )3) View from bottom: The sum of white pixels when scanning an input image vertically from the bottom (y2) up (y3) until touching a black pixel.3) Top view: The sum of white pixels when scan an input image vertically from top (y0) to bottom (y1) until it touches a black pixel. Vertical Profile= (top view)/(bottom view) Input Image Size (128 128) left= 0 right= 582HP= 0Image size (256 horizontal profile are scale invariant8. Zone Density (ZD) Characteristics • The density of each zone is obtained by dividing the number of foreground pixels in each zone by the total number of pixels in each zone.16 zones with normalized Devnagari number 32*32 „3‟.