Why are simple texture detector operators eg standard deviat

Why are simple texture detector operators e.g. standard deviation of pixels) not particularly effective when differentiating between a large number of different textures?

Solution

texture analysis is mainly concentrated on the neighbourhood pixels. as the neighbour pixels have a great intensity changes it cannot recoginize efficiently because of its statistical comparison. so the imple texture detector operators not particularly effective when differentiating between a large number of different textures?