Compares two shapes

result = cv.matchShapes(contour1, contour2)
result = cv.matchShapes(..., 'OptionName', optionValue, ...)

Input

• contour1 First contour or grayscale image. If it's contour, a cell array of 2-element vectors (of int32 or single type) of the form {[x,y], ...}. In case of image, a single-channel 8-bit or floating-point 2D matrix.
• contour2 Second contour or grayscale image, with a similar type as the first contour.

Output

• result Output score.

Options

• Method Comparison method, default 'I1'. One of:
  • I1
  • I2
  • I3
• Parameter Method-specific parameter (not supported now). default 0

The function compares two shapes. All three implemented methods use the Hu invariants (see cv.HuMoments).

The following shape matching methods are available:

• I1:

  I1(A,B) = \sum{i=1,...,7} | 1/mi^A - 1/mi^B |

• I2:

  I2(A,B) = \sum{i=1,...,7} | mi^A - mi^B |

• I3:

  I3(A,B) = \max{i=1,...,7} | mi^A - mi^B | / | m_i^A |

where A denotes contour1, B denotes contour2, and:

m_i^A = sign(h_i^A) * log(h_i^A)
m_i^B = sign(h_i^B) * log(h_i^B)

and h_i^A, h_i^B are the Hu moments of A and B, respectively.