Method for realizing classification of scene images

Abstract

The invention discloses a method for realizing classification of scene images. In the method, the scene images are classified by adopting two stages of classifiers, wherein the first-stage classifier obtains candidate classifications by utilizing global structural information characteristics, and judges similar classification pairs according to classification results; and the second-stage classifier distinguishes similar classifications by utilizing local texture information characteristics. By adopting the cascade of the classifiers and comprehensively utilizing the global structural information characteristics and the local texture information characteristics of the scene images, the method can realize robust classification of different scene classifications and effectively distinguish similar scene classifications.

Description

technical field The invention belongs to the technical field of pattern recognition and information processing, and relates to the automatic processing technology of digital images, in particular to a classification method of scene images. Background technique With the wide application of various image acquisition devices such as digital cameras, cameras, ultra-high-speed scanners, and the rapid development of the Internet, the number of digital images has grown exponentially. According to incomplete statistics, more than 18 billion digital images were generated in 2004 alone. images, and Google Image Search has indexed hundreds of millions of images, so how to effectively classify images is becoming more and more important. A person can usually recognize more than 10,000 categories of visual objects, and the recognition process is fast and effortless. It is also robust to perspective, light, occlusion and background confusion, and it only takes a few minutes to recognize a ...

AI Technical Summary

Problems solved by technology

The existing scene image classification methods often use a single classifier, and only select a single-stage classifier to complete the final classification according to the empirical observation of the scene image. This method cannot effectively use various feature information of the image, and ignores the The fusion and cascading of classifiers can take advantage of the complementary advantages of different features, so that the classification of scene images has not yet achieved satisfactory results

Images