A novel approach of image retrieval towards segmentation. (16th June 2021)

• Record Type: Journal Article
• Authors: Tamilkodi, R.; Nesakumari, G. Rosline
• Is Part Of: International journal of image mining. Volume 4:Number 1(2020)
• Page Start: 46
• Page End: 58
• Publication Date: 2021-06-16
• Subjects: segmentation -- retrieval -- feature extraction; Image processing -- Periodicals; Data mining -- Periodicals; 006.42
• Languages: English
• ISSNs: 2055-6039

Abstract

At the present time, it becomes simple to stockpile vast quantity of images by means of image processing approach. The hasty right of entry to these ample groups of images and retrieve related images of a specified image from this massive assortment of images presents key challenges and requires competent algorithms. In this manuscript, the authors intend a structure which is proficient to select the most appropriate features to examine newly expected images in that way improving the retrieval accuracy and competence. An enhanced algorithm is projected now. The algorithm comprises of scheming feature vectors after segmentation which will be used in likeness contrast linking query image and database images. The structure is skilled for dissimilar images in the database. The projected method has been tested on a variety of real images and its concert is found to be reasonably acceptable when compared with the concert of conventional methods of content-based image retrieval. The major objective of the anticipated method is to endow with an exact outcome with lesser computational time.