Relaxation labeling for non-rigid point matching under neighbor preserving

Non-rigid point matching has received more and more attention.Recently,many works have been developed to discover global relationships in the point set which is treated as an instance of a joint distribution.However,the local relationship among neighboring points is more effective under non-rigid transformations.Thus,a new algorithm taking advantage of shape context and relaxation labeling technique,called SC-RL,is proposed for non-rigid point matching.It is a strategy that joints estimation for correspondences as well as the transformation.In this work,correspondence assignment is treated as a soft-assign process in which the matching probability is updated by relaxation labeling technique with a newly defined compatibility coefficient.The compatibility coefficient is one or zero depending on whether neighboring points preserving their relative position in a local coordinate system.The comparative analysis has been performed against four state-of-the-art algorithms including SC,ICP,TPS-RPM and RPM-LNS,and the results denote that SC-RL performs better in the presence of deformations,outliers and noise.

Label-noise robust classification with multi-view learning

Label noise is often contained in the training data due to various human factors or measurement errors,which significantly causes a negative effect on classifiers.Despite many previous methods that have been proposed to learn robust classifiers,they are mainly based on the single-view feature.On the other hand,although existing multi-view classification methods benefit from the more comprehensive information,they rarely consider label noise.In this paper,we propose a novel label-noise robust classification model with multi-view learning to overcome these limitations.In the proposed model,not only the classifier learning but also the label-noise removal can benefit from the multi-view information.Specifically,we relax the label matrix of the basic multi-view least squares regression model,and develop a nonlinear transformation with a natural probabilistic approximation in the process of labels,which is conveniently optimized and beneficial to improve the discriminative ability of classifiers.Moreover,we preserve the intrinsic manifold structure of multi-view data on the relaxed label matrix,facilitating the process of label relaxation.For optimizing the proposed model with the nonlinear transformation,we derive a lemma about the partial derivation of the softmax related function,and develop an efficient alternating algorithm.Experimental evaluations on six real-world datasets confirm the advantages of the proposed method,compared to the related state-of-the-art methods.