An Image Classification Method Based on Saliency-Guided Unsupervised Feature Learning

Abstract

The invention discloses an image classification method based on saliency-guided non-supervised feature learning, which belongs to the field of machine learning and computer vision. The image classification method includes saliency-guided pixel point acquisition, unsupervised feature learning, image convolution, local contrast normalization, spatial pyramid pooling, fusion central prior and image classification. Using this classification method, the saliency detection is used to collect representative pixels in the image data set, and the representative pixels are trained through sparse autoencoder, an unsupervised feature learning method, to obtain high-quality image features. Obtain the features of training samples and test sets through image convolution operations, perform local contrast normalization and spatial pyramid pooling on the convolutional features, fuse the pooled features with the central prior features, and use the liblinear classifier to classify the image sort. This method can obtain efficient and robust image features, which can significantly improve the classification accuracy of multi-class images.

Description

technical field [0001] The invention relates to an image classification method based on saliency-guided non-supervised feature learning, which belongs to the technical field of machine learning and computer vision. Background technique [0002] With the development of multimedia technology, image classification has become the focus of research in the field of computer vision. Image classification is to divide images into different preset categories according to certain attributes of images. How to effectively express images is the key to improving the The key to the accuracy of image classification, the selection and extraction of features are currently difficult problems in image classification. With the rapid development of the mobile Internet, human society has entered the era of big data. Although traditional feature learning methods such as SIFT and HOG can extract some features of images, they have also achieved good results in image classification, but this artificia...

AI Technical Summary

Problems solved by technology

Although traditional feature learning such as SIFT and HOG can extract some features of images, they have also achieved good results in image classification, but this artificial design feature method has certain defects.

However, the traditional supervised feature learning method, by learning manual labeling data, is outdated in the era of big data.

Images