Cross Reference Strategies for Cooperative Modalities

In this paper, we propose a Generic video visual approach for find the best cluster video shots. It is one kind of categorization technique for classifying clusters. New Clusters Sequences generation or creation using local videos of sequence and global temporal videos sequences. Previously we discuss about relevance feedback, pseudo relevance feedback, Meta search strategy, Content based video retrieval are not provides better ranking features. Some kind’s videos are gets the wrong ranking features and wrong clusters also. These clusters are works as missing features clusters. Now we introduce perfect clusters creation with genetic algorithm. We take major input as a different or multiple numbers of modalities. Each and every model of genomes we recognize same like collects the different video of different genomes. It is good classification and categorization process. Classification works matrix based features selection from multiple modalities. It is the regroup and new video creation. New video contains good density and content features. This is optimal cluster creation.