Dictionary based automatic target recognition

There were different methods were developed for moving and stationary target acquisition and recognition (MSTAR) such as Non Negative Matrix Approximation (NNMA), Principal Component analysis (PCA), Linear Discriminant Analysis (LDA). But they have their lagging in the accuracy level. This paper explains that the recognition of multi view target automatically. To recognise such target without knowing the pose, a method used named Content Based Image Retrieval. In this method we can use the correlation of multiple views for the same target. It provides a greater accuracy for automatic target recognition (ATR) and convenient for different angles and configurations. It can out the neighbour of the new test image which is not found in the dictionary.