Image retrieval method based on cgci-sift local features

Abstract

The invention discloses an image search method based on a CGCI-SIFT (consistence index-scale invariant feature transform) partial feature. The image search method is realized based on CGCI-SIFT and comprises the following steps of: starting from the strength and the distribution of the influence of a neighborhood domain pixel to a key point; establishing a periphery partial feature descriptor through gray level texture comparison strength information; and subsequently establishing a central partial descriptor by combining direction gradient information having relatively high central feature point description so as to form a final description, wherein the CGCI-SIFT utilizes the contrast property of a partial area and is combined with gradient information of the original SIFT algorithm, rather than that the SIFT in which the weight and the direction of the gradient are singly stored, so that the CGCI-SIFT has relatively comprehensive geometric and optical conversion invariance. Due to the utilization of the gray level texture comparison strength information, the CGCI-SIFT is simple in calculation, thereby being relatively efficient and relatively applicable to real-time application. Tests show that according to the search method, the performance is stable, the search time is short, and a remarkable improvement in the search effect can be realized.

Description

technical field [0001] The invention relates to an image retrieval method, in particular to a feature-based image retrieval method, in particular to a method for image retrieval using a new local feature description information, which belongs to the field of image retrieval. Background technique [0002] The local feature description method is a very stable method in image matching. The local feature description method has low redundancy, no need to segment the image in advance, and has invariance under multiple transformations, which makes the local feature description method widely used, such as content-based large-scale retrieval, video analysis, repeated detection, object Identification, travel photography, 3D reconstruction and retrieval, etc. [0003] A good local image feature method should have the following characteristics: no need to pre-segment the image, high repetition rate of feature detection, low dimensionality of feature descriptors, easy to achieve fast ma...

AI Technical Summary

Problems solved by technology

[0004] However, there are various problems in the existing local feature description methods, such as the high dimensionality of SIFT descriptors and the simple use of gradient histograms to construct descriptors, which limits its application in real-time and some special fields.

Images