Class: KvgCharacterRecognition::Preprocessor

Inherits:
Object
  • Object
show all
Includes:
Normalization
Defined in:
lib/kvg_character_recognition/preprocessor.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from Normalization

#bi_moment_normalize, #line_density_normalize, #point_density_normalize

Constructor Details

#initialize(interpolate_distance, size, smooth = true, smooth_weights = [1,2,3,2,1]) ⇒ Preprocessor 



5
6
7
8
9
10
11
 
# File 'lib/kvg_character_recognition/preprocessor.rb', line 5

def initialize interpolate_distance, size, smooth=true, smooth_weights=[1,2,3,2,1]
  @smooth = smooth
  @smooth_weights = smooth_weights
  #@downsample_rate = downsample_rate
  @interpolate_distance = interpolate_distance
  @size = size
end

Instance Attribute Details

#interpolate_distance ⇒ Object

Returns the value of attribute interpolate_distance.



4
5
6
 
# File 'lib/kvg_character_recognition/preprocessor.rb', line 4

def interpolate_distance
  @interpolate_distance
end

#number_of_points ⇒ Object

Returns the value of attribute number_of_points.



4
5
6
 
# File 'lib/kvg_character_recognition/preprocessor.rb', line 4

def number_of_points
  @number_of_points
end

#size ⇒ Object

Returns the value of attribute size.



4
5
6
 
# File 'lib/kvg_character_recognition/preprocessor.rb', line 4

def size
  @size
end

#smooth(stroke) ⇒ Object

A simple smooth method using the following formula p’(i) = (w(-M)*p(i-M) + … + w(0)*p(i) + … + w(M)*p(i+M)) / S where the smoothed point is a weighted average of its adjacent points. Only the user input should be smoothed, it is not necessary for kvg data. Params:

stroke

array of points i.e [[x1, y1], [x2, y2] …]



27
28
29
 
# File 'lib/kvg_character_recognition/preprocessor.rb', line 27

def smooth
  @smooth
end

#smooth_weights ⇒ Object

Returns the value of attribute smooth_weights.



4
5
6
 
# File 'lib/kvg_character_recognition/preprocessor.rb', line 4

def smooth_weights
  @smooth_weights
end

Instance Method Details

#interpolate(stroke) ⇒ Object

This method interpolates points into a stroke with given distance The algorithm is taken from the paper preprocessing techniques for online character recognition



50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
 
# File 'lib/kvg_character_recognition/preprocessor.rb', line 50

def interpolate stroke
  current = stroke.first
  new_stroke = [current]

  index = 1
  last_index = 0
  while index < stroke.length do
    point = stroke[index]

    #only consider point with greater than d distance to current point
    if Math.euclidean_distance(current, point) < @interpolate_distance
      index += 1
    else

      #calculate new point coordinate
      new_point = []
      if point[0].round(2) == current[0].round(2) # x2 == x1
        if point[1] > current[1] # y2 > y1
          new_point = [current[0], current[1] + @interpolate_distance]
        else # y2 < y1
          new_point = [current[0], current[1] - @interpolate_distance]
        end
      else # x2 != x1
        slope = (point[1] - current[1]) / (point[0] - current[0]).to_f
        if point[0] > current[0] # x2 > x1
          new_point[0] = current[0] + Math.sqrt(@interpolate_distance**2 / (slope**2 + 1))
        else # x2 < x1
          new_point[0] = current[0] - Math.sqrt(@interpolate_distance**2 / (slope**2 + 1))
        end
        new_point[1] = slope * new_point[0] + point[1] - (slope * point[0])
      end

      new_point = new_point.map{ |num| num.round(2) }
      if current != new_point
        new_stroke << new_point

        current = new_point
      end
      last_index += ((index - last_index) / 2).floor
      index = last_index + 1
    end
  end

  new_stroke
end

#preprocess(strokes) ⇒ Object

preprocess steps bi moment size normalization, smooth and interpolate



14
15
16
17
18
19
 
# File 'lib/kvg_character_recognition/preprocessor.rb', line 14

def preprocess strokes
  bi_moment_normalize(strokes).map do |stroke|
    stroke = smooth(stroke) if @smooth
    smooth(interpolate(stroke))
  end
end