Retrieval method of color images based on quaternion invariants

Abstract

The invention discloses a retrieval method of color images based on quaternion invariants. The method includes expanding PHT moments according to the principle that quaternions contain color information, expanding the PHT moments to quaternion level, acquiring a new image-describing feature vector according to the character that the PHT moments can describe shape features of images, and designing a new retrieval system of color images based on quaternion invariants. Experimental results show that the system has such features that the system can effectively represent the colors and shape features of the images, the design is simple, implementing is easy, and image feature extraction needs no manual annotation; the manual annotation process tedious and error prone is avoided, and retrieval precision and efficiency is greatly improved.

Description

technical field [0001] The invention belongs to the field of image processing, in particular to a color image retrieval method based on quaternion invariants that can effectively improve retrieval precision and efficiency. Background technique [0002] Text-based image retrieval technology (TBIR) follows the traditional text retrieval technology. It does not consider the inherent color, texture, shape and other content characteristics of the image itself, but uses keywords to describe the image. Retrieve related images in the form of . This technology has the following two defects: firstly, due to the continuous expansion of the current image database, it takes a lot of time and manpower to manually label each image in the database; In addition, the process of manually selecting keywords will contain strong subjectivity, which may cause deviations in image understanding and directly affect the retrieval effect of images. In order to overcome the difficulties brought by tex...

AI Technical Summary

Problems solved by technology

This technology has the following two defects: firstly, due to the continuous expansion of the image database, it takes a lot of time and manpower to manually label each image in the database; Connotation, and the process of manually selecting keywords will contain strong subjectivity, which may cause deviations in image understanding and directly affect the retrieval effect of images

[0003] In image visual features, color and shape are the two most important features. However, the existing image feature extraction methods often deal with the two separately. There are many retrieval methods for color features, such as based on traditional histograms, color Correlogram, color matrix, spatial distribution method based on main color, etc.; contour-based and region-based retrieval methods for shape features, such as wavelet transform Gaussian distribution method, feature moment (Zernike moment, Legendre moment, etc.) method, etc., However, the retrieval efficiency of individual color features or trait features is low, which is far from meeting the requirements of image retrieval efficiency.

Images