Palmprint feature extraction using weight coding based non-negative sparse coding

A novel palmprint feature extraction method is proposed by using the Weight Coding Based Non-negative Sparse Coding (WCBNNSC). The WCBNNSC algorithm can model the respective field of V1 in the primary visual system of brain. And this algorithm includes more image information than the early Non-negative Sparse Coding (NNSC). Utilizing the WCBNNSC algorithm, the feature basis vectors of palmprint images can be successfully learned. These features behave locality, orientation, and spatial selection, which is similar to the respective field feature of V1 in visual cortex. Further, using the features extracted, the palmprint reconstruction task can be successfully implemented. Moreover, compared with other palmprint feature extraction methods, simulation results show that our method proposed here is indeed efficient and useful in performing the feature extraction task of palmprint images.