Hamming distance between two regions using normalization.
The operator hamming_distance_norm returns the hamming distance between two regions, i.e., the number of pixels of the regions which are different (Distance). Before calculating the difference the region in Regions1 is normalized onto the regions in Regions2. The result is the number of pixels contained in one region but not in the other:
Distance = |Norm(Regions1) intersection ~Regions2| +
|Regions2 intersection ~Norm(Regions1)|
The parameter Similarity describes the similarity between the
two regions based on the hamming distance Distance:
Similarity = ( 1 - Distance ) / ( |Norm(Regions1)| + |Regions2| )
The following types of normalization are available:
'center': The region is moved so that both regions have the save center of gravity.
If both regions are empty Similarity is set to 0. The regions with the same index from both input parameters are always compared.
In both input parameters the same number of regions must be passed.
|
Regions1 (input_object) |
region(-array) -> object |
| Regions to be examined. | |
|
Regions2 (input_object) |
region(-array) -> object |
| Comparative regions. | |
|
Norm (input_control) |
string(-array) -> string |
| Type of normalization. | |
| Default value: 'center' | |
| List of values: 'center' | |
|
Distance (output_control) |
integer(-array) -> integer |
| Hamming distance of two regions. | |
| Assertion: Distance >= 0 | |
|
Similarity (output_control) |
real(-array) -> real |
| Similarity of two regions. | |
| Assertion: (0 <= Similarity) && (Similarity <= 1) | |
If F is the area of a region the mean runtime complexity is O(sqrt(F)).
hamming_distance_norm returns the value 2 (H_MSG_TRUE) if the number of objects in both parameters is the same and is not 0. The behavior in case of empty input (no input objects available) is set via the operator set_system('no_object_result',<Result>). The behavior in case of empty region (the region is the empty set) is set via set_system('empty_region_result',<Result>). If necessary an exception handling is raised.
hamming_distance_norm is reentrant and automatically parallelized (on tuple level).
threshold, regiongrowing, connection
intersection, complement, area_center
Region processing
