Figure 13From: A neural computational model for bottom-up attention with invariant and overcomplete representationSome examples of invariant feature descriptors obtained by pooling and then refining. Each row lists a descriptor Ω j in which orientations and frequencies are similar but phases are different.Back to article page