Multiscale Feature Extraction For Content Based Image Retrieval Using Gabor Local Tetra Pattern

In this project the image is retrieved using Local Tetra Pattern (LTrP) for Content Based Image Retrieval (CBIR). It gives the path to retrieve the needed information based on the image content. The earlier version of CBIR was based on Local Binary Pattern, Local Derivative Pattern and Local Ternary Pattern. These methods extract information based on the distribution of edges which are coded using only two directions. The performance of these methods is little less and thus it can be improved by differentiating the edges in more than two directions. So we propose local tetra pattern, in this we encodes the relationship between the referenced pixel and its neighbours based on the directions that can be calculated using second order derivatives in horizontal and vertical directions. The effectiveness of our proposed algorithm can be analyzed by combining it with the Gabor Transform. The performance of the proposed method is compared with the LBP, the LDP and the LTP based on the results obtained using benchmark image databases.