期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2015
卷号：6
期号：6
页码：5198-5203
出版社：TechScience Publications
摘要：The data retrieval process needs a storage and search methodology to find the data according to the query inputs. The presented work is focused on the image data retrieval using the query inputs. There are two different kinds of input are frequently used for image search namely text or images. But the text query can be misguiding the search outcomes and search process but the query by image can be used for accurate data identification. Therefore the content based image retrieval technique is proposed for study using the image query inputs. The proposed content based image retrieval process need to identify the image contents using their valuable features by which the image and their properties. Therefore the image contents are evaluated using the edge histogram for shape feature extraction, grid color movement analysis is performed for color feature evaluation and the local binary pattern is estimated for texture analysis. These features are normalized and stored for extraction of image. Therefore the query image is also extracted for comparison and accurate image extraction. But for enhancing more the multi-scale saliency feature is also used with the feature extraction. These features are help to identify the image groups by which the noisy images are separated and only the good images are considered for image comparison and extraction. Additionally that features also helps to improve the time and space complexity by reducing the number of comparison during the KNN based images search. The implementation of the proposed technique is performed using JAVA technology and their performance is evaluated in different performance parameters i.e. precision, recall, fmeasures, time complexity, and space complexity. The obtained performance of the system shows the improved outcomes and accurate image retrieval as described in the query image.
关键词：image retrieval; multi-scale saliency model;content based image analysis; feature extraction;implementation
