Content Based Image Retrieval Using Multi-Sequential Search
Sanchit Shinde, Mrudula Desai, Mihika Gupte, Divya Racha · 2018
In Content-Based Image Retrieval (CBIR) the indexing of the image and the representation is done using visual contents of an image like color, shape, texture and spatial layout. The on-going development and research in CBIR are mainly focused on the improvement of various methods for examining, cataloging and indexing image databases. Along with this development, several efforts are also being made in evaluating the performance of the image retrieval systems. Depending on the method incorporated to generate the similarity measure for the comparisons of different features and the vectors the accuracy and response of the system would vary. The proposed system is exclusive as it takes into consideration one particular feature at one step and uses the result of the above prior step as the input for the next step instead of fusing all the features at one step and thus providing an edge over the earlier used methods and also showing improvement in the retrieval quality. In the proposed system the accuracy of color histogram-based matching can be increased by using Color Coherence Vector. Instead of taking into consideration the exact shape, the approximation of the boundary is considered in order to enhance the speed of shape-based matching.