AN IMAGE RETRIEVAL SYSTEM BASED ON SIGNIFICANCE FEEDBACK

Abstract:

This paper presents a picture reclamation scheme for versatile images by “Search by query image “model. Firstly, the images are considered, and denoising is done later the image data is analysed using PCA. Now, the images are classified using the AdaBoost classifier. The irrelevant images are filtered out there by reducing the search space and computational time. The images are categorized into various categories. For more promising results the relevance feedback is considered.

Further the images are considered for similarity matching process. The images are stored in the storage and various low-level features such as texture colour, shape and pattern. The various low level – features are extracted using histograms, various filtering and edge detection techniques. In this study, we also retrieve resulting images basing on the various low-level features and also a fusion of features from the query image such as colour and shape.

Keywords:

PCA, CBIR, RF, AdaBoost, SVM