Image classification method based on cascaded codebook generation

Abstract

The invention provides an image classification method based on cascaded codebook generation. In the method, a cascaded codebook generation means is adopted, and a sample mapping space is dynamically adjusted according to the change of training samples, so that the training samples are easier to separate to train a classifier based on the current codebook. A test image sequentially passes through each trained classifier; when outputs of all classifiers are positive, a category tab of the image is judged to be positive; or the category tab of the image is judged to be negative. As a cascaded codebook generation method is adopted in the method, the number of dimensions represented by the image is reduced while each turn of codebooks keep stronger discrimination, and the conflict between the diversity of the codebook and the dimension disaster is solved; meanwhile, as the cascaded codebook generation means is adopted in the method, the number of positive samples and the number of negative samples trained in each turn are the same, and the common problem of nonuniform samples in the classification of special categories is ingeniously solved; moreover, the image classification method has an important application value.

Description

technical field [0001] The invention belongs to the technical field of pattern recognition and information processing, and relates to an image classification method based on cascade codebook generation, in particular to an image classification method. Background technique [0002] Due to the low price of digital products such as digital cameras, pictures and video data can be easily taken and stored in an electronic form that is convenient for computer processing. At the same time, the rapid development of Internet resources makes the majority of users begin to face a huge data resource. It has become unrealistic to simply use manpower to maintain and organize these data. Therefore, developing a technology to effectively organize these personal or corporate picture and video data has become a hot issue. However, computers cannot "see" very clearly: they cannot convert the color pixels in an image into a higher-level semantic representation like humans, and then complete th...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the difficulties existing in the multi-class classification problem in the prior art, the purpose of the present invention is to provide a method that can not only gather discriminative local salient area image blocks between classes as a codebook, but also effectively reduce the classification vector Solution to the Curse of Dimensionality Problem of Representation

At the same time avoid the problem of sample imbalance

Images