Calculate gray value features from a co-occurrence matrix.
The procedure calculates from a co-occurence matrix (CoocMatrix) the energy (Energy), correlation (Correlation), local homogeneity (Homogeneity) and contrast (Contrast).
The operator cooc_feature_matrix calculates the gray value features from the part of the input matrix generated by gen_cooc_matrix corresponding to the direction matrix indicated by the parameters LdGray and Direction according to the following formulae:
Energy:
width width
---- ----
\ \
Energy = / / cij * cij
---- ----
i=0 j=0
(Measure for image homogeneity)
Correlation:
width width
---- ----
\ \
/ / (i - ux) * (j - uy) * cij
---- ----
i=0 j=0
Correlation = -----------------------------------------
sx * sy
(Measure for gray value dependencies)
Local homogeneity:
width width
---- ----
\ \ 1
Homogeneity = / / ------------- * cij
---- ---- 1 + (i - j)^2
i=0 j=0
Contrast:
width width
---- ----
\ \
Contrast = / / (i - j)^2 * cij
---- ----
i=0 j=0
(Measure for the size of the intensity differences)
where
width 2^LdGray - 1
cij Entry of co-occurrence matrix
width width
---- ----
\ \
ux / / i * cij
---- ----
i=0 j=0
width width
---- ----
\ \
uy / / j * cij
---- ----
i=0 j=0
width width
---- ----
\ \
sx^2 / / (i - ux)^2 * cij
---- ----
i=0 j=0
width width
---- ----
\ \
sy^2 / / (i - uy)^2 * cij
---- ----
i=0 j=0
The region of the input image is disregarded.
|
CoocMatrix (input_object) |
image -> object : real |
| Co-occurrence matrix. | |
|
Energy (output_control) |
real -> real |
| Homogeneity of the gray values. | |
|
Correlation (output_control) |
real -> real |
| Correlation of gray values. | |
|
Homogeneity (output_control) |
real -> real |
| Local homogeneity of gray values. | |
|
Contrast (output_control) |
real -> real |
| Gray value contrast. | |
The operator cooc_feature_matrix returns the value 2 (H_MSG_TRUE) if an image with defined gray values is passed and the parameters are correct. The behavior in case of empty input (no input images available) is set via the operator set_system(::'no_object_result',<Result>:). If necessary an exception handling is raised.
cooc_feature_matrix is reentrant and automatically parallelized (on tuple level).
intensity, min_max_gray, entropy_gray, select_gray
Foundation
