Hierarchical Sparse Method With Applications in Vision and Speech Recognition

Abstract

A new approach for feature extraction using neural response has been developed in this paper through combining the hierarchical architectures with the sparse coding technique. As far as proposed layered model, at each layer of hierarchy, it concerned two components that were used are sparse coding and pooling operation. While the sparse coding was used to solve increasingly complex sparse feature representations, the pooling operation by comparing sparse outputs was used to measure the match between a stored prototype and the input sub-image. It is recommended that value of the best matching should be kept and discarding the others. The proposed model is implemented and tested taking into account two ranges of recognition tasks i.e. image recognition and speech recognition (on isolated word vocabulary). Experimental results with various parameters demonstrate that proposed scheme leads to extract more efficient features than other methods.

Using Derived kernel as a new Method for Recognition a Similarity Learning.

“…The importance of this connection rely on the fact where function is capable of mapping a set of feature paths from the lower patches to a solitary feature path at the upper patches. Reader is recommended to consult [24,25,28] for further detail treatment.…”

“…The following is a brief description of the calculated framework, shown in [24,25,27], using the hierarchical model of the construction, every input here is consisted of segments of growing size.…”

“…Consequently, it is notable that only at the phase of selecting pattern, there is a variation then the remaining steps remain same as [24,25].…”

“…In [11], Hong Li et al suggested a pattern selection technique for the similarity measure capability of the derived kernel based neural response, when [23] presenting a hierarchical feature extraction method called local neural response (LNR) to recognize. Another technique is suggested in [24,25] to a feature extraction scheme using neural response by linking the hierarchical architectures with the sparse coding. An image classification algorithm as a semi-supervised sparse based neural response was examined in [26].…”