Action model prediction and analysis for CBMR application

Content base multimedia retrieval (CBMR) has been used in the context of information retrieval for various real time practical usages. The CBMR model process on the feature extracted and classify the observation based on training and testing process. Wherein, feature extraction are performed to define the detail content of a observing sample, the recurrent feature values increase the overhead. This paper define a new approach of low dimensional feature representation and feature grouping for content information retrieval based on energy feature interpolation and multi linear mapping. The approach defines a weighted cluster grouping to minimize the classification overhead. The proposed system analyze the proposed solution for a k-fold test analysis for Weizmann datasets.