An Integrated, Optimum Feature Extraction and Evolutionary Computation Approach for the CBIR

The visual information exploited in the 21^st century generated from the advanced digitization techniques is playing the significant role in knowledge communication and transformation. The Governments, Academics, as well as the Corporate sectors, are daily relying on such huge volume of the digital knowledge in the form of images. The robust, accurate and reliable retrieval technique for the specific image is still the great challenge in front of the researchers across the world. This paper proposes an image retrieval algorithm that communicates with the user. The user of this system decides the end results and based on his/her satisfaction. The visual features are also selected by the user. The system helps the user in obtaining an optimum result by using a set of Artificial Neural Network based classifiers. These classifiers try to maximize helps the system to maximize the satisfaction of the users. An Evolutionary Computational method like the Genetic Algorithm is efficiently and effectively deployed here, that optimizes the retrieval process. It is used as the Relevance Feedback in the system. The algorithm is tested on the huge collection of diversified Corel Image. Concatenated form of the visual features, classify the images efficiently. Visual features such as multi-resolution color, Local Binary Pattern (LBP), Histogram Oriented Gradient (HOG), Scale Invariant Feature Transform (SIFT), Edge Histogram and texture features are used here to represent an image digitally. The experimental result shows that the performance of this Image Retrieval system is qualitatively improved than the existing similar algorithms.