Content Retrieval using Hybrid Feature extraction from Query Image
Nehal M. Varma, Arshi Riyaz · 2018
The method of retrieval of similar images on the basis of various contents of query images finds its application in various domains such as Identification of Fingerprints, Information Systems of Biodiversity, Digitally acquainted Libraries, Prevention of crimes, Medical, Research of Historical places & many others. With the vast growth in different types of images over internet, it becomes a challenging task to retrieve efficient result from the large database. Content based image retrieval (CBIR) gives solution to the stated problem as it uses visual image contents or features such as color, shape, texture unlike only focusing on keywords deals with particular image. This work proposes a mixture of hybrid features for based image retrieval in which color features are obtained using color histogram method which will ultimately give the color gradient. The outer and inner edge would be extracted through gabor wavelet method. The final features vector would be constructed by combining these above methods to retrieve visually similar images. The developed algorithm has been tested on wide range of datasets and the performance is recorded in terms of matching score and subjective analysis. The obtained results are competitive with respect to recently proposed methodologies.