PROJECT TITLE :

Sparsity-Induced Similarity Measure and Its Applications

ABSTRACT:

The structures of feature vectors-based semisupervised/supervised learning have gained considerable interest in recent times thanks to their effectiveness for better object modeling and classification. In many Machine Learning and pc vision tasks, a critical issue is that the similarity between two feature vectors. During this paper, we tend to present a unique technique to measure similarities among feature vectors by decomposing each feature vector as an ℓone sparse linear combination of the rest of the feature vectors. The main plan is that the coefficients in such sparse decomposition reflect the features' neighborhood structure, so providing higher similarity measures among the decomposed feature vector and the remainder of the feature vectors. The proposed approach is applied to label propagation and action recognition, and is evaluated on many commonly used datasets. The experimental results show that the proposed sparsity-induced similarity live considerably improves the performance of both label propagation and action recognition.