Pooling-Invariant Image Feature Learning

Unsupervised dictionary learning has been a key com-ponent in state-of-the-art computer vision recognition ar-chitectures. While highly effective methods exist for patch-based dictionary learning, these methods may learn redun-dant features after the pooling stage in a given early vi-sion architecture. In this paper, we offer a novel dictionary learning scheme to efficiently take into account the invari-ance of learned features after the spatial pooling stage. The algorithm is built on simple clustering, and thus enjoys ef-ficiency and scalability. We discuss the underlying mecha-nism that justifies the use of clustering algorithms, and em-pirically show that the algorithm finds better dictionaries than patch-based methods with the same dictionary size. 1