Object detection training process does not support a polygon area with the same size of the input image
Mantis Issue 252, reported by mgrizonnet, assigned to ghost, created: 2011-03-11
Some tests of the object detection chains (activated with the OTB_OBJECT_DETECTION_PERFORMANCES_TESTING cmake option) fail because the input polygons (Working areas) have the same size of the input image. There must be a problem in the LabeledSampleLocalizationGenerator class
I close the bug. The commit http://hg.orfeo-toolbox.org/OTB/rev/755a24d1cd45 corrects it.
NULL