An Unsupervised Nighttime Image Classification Method Based on Feature Augmentation

Abstract

The invention belongs to the technical field of computer vision recognition, and relates to an unsupervised night image classification method based on feature augmentation. The classification network is trained using the public dataset with daytime image classification labels, the feature vector of the input image is extracted by the classification network, and the feature mean and covariance matrix of each category are calculated; the unlabeled nighttime image is input into the classification network to obtain the pseudo label, calculate the feature mean and covariance matrix of each category of nighttime images in the feature space according to the pseudo-label; perform weighted average of the covariance matrices obtained from daytime and nighttime images of the same category to obtain the final covariance matrix; according to the feature mean of each category of nighttime images Perform feature sampling with the weighted average covariance matrix; retrain the classification network with the sampled eigenvalues ​​and the original eigenvalues. By learning the feature distribution of the labeled daytime images, the invention augments the nighttime data at the feature level, thereby realizing the unsupervised classification of the nighttime images.

Description

technical field [0001] The invention belongs to the technical field of computer vision recognition, in particular to an unsupervised nighttime image classification method based on feature amplification. Background technique [0002] Image classification is the most classic task in the field of computer vision recognition, and it is also the basis of many other visual problems, which has great practical value and application prospects. Image classification is essentially a pattern classification problem, and its goal is to divide different images into different categories to achieve the smallest classification error. With the success of convolutional neural network (CNN), deep learning has been proven to be an effective solution to image classification problems. [0003] The existing large-scale public datasets involving image classification mainly include ImageNet, COCO, PascalVOC, etc. However, these datasets are basically images collected in daytime environments. Studies...

AI Technical Summary

Problems solved by technology

However, the nighttime images generated by this method do not conform to the real data distribution, and there is also a field gap with the real nighttime data

Images