Spatial Fisher vector based image classification method

Abstract

The invention discloses a spatial Fisher vector based image classification method which mainly solves the problems of 'foreground and background visual ambiguities' of images and lack of feature point space distribution information in image description vectors in the prior art. The method includes the implementation steps: (1) extracting 'scale-invariant feature transformation' feature points of all images; (2) in a feature point space of images in a training set, clustering the feature points by the aid of a Gaussian hybrid clustering algorithm to obtain a code book; (3) utilizing gradient vectors and Cholesky components for generating a Fisher vector of each image; (4) performing 2X2 space region division on each image, and making statistics on number of feature points and coordinates of each cell block; (5) utilizing a Fisher vector of each cell block for stitching to generate a space Fisher vector of each image; and (6) utilizing a support vector mechanism for constructing a classification hyperplane to achieve image classification. The method has the advantages that image information can be described accurately, accuracy rate of image classification is improved, and the method can be used for large-scale image classification and retrieval system construction.

Description

technical field [0001] The invention relates to the technical field of image processing, and relates to an image classification method, which can be used for intelligent image classification management and web image push. Background technique [0002] Image classification is a typical problem in the field of computer vision, and it is particularly prominent with the massive growth of multimedia data. Image classification is usually based on the semantic content of the image, such as specific scenes, specific inclusions, etc., to add different category labels to the image to achieve image classification. Images are often affected by imaging factors such as viewing angle, illumination, and occlusion, which brings great challenges to image classification. [0003] At present, image classification methods mainly include image classification methods based on text labels and image classification methods based on content. [0004] Image classification methods based on text labels...

AI Technical Summary

Problems solved by technology

The main shortcomings of this patent are: first, when assigning feature points to each word in the visual dictionary, the hard allocation method of "one feature point corresponds to one word" is used, and one feature point can only be assigned to On one codeword, the "foreground and background visual ambiguity" phenomenon in real images is ignored; second, the frequency histogram constructed during image vectorization is actually an unordered collection of local feature points on the codebook , does not consider the spatial layout information of the feature points at all. For images with obvious hierarchical characteristics, the spatial layout information of the feature points may contain information that is discriminative for classification. Ignoring this information will result in the generated frequency histogram Each frequency component in is completely disordered, and it is impossible to describe the position range of the corresponding object in the image in the image space

Images