Published Paper Details:

ABSTRACT The traditional unsupervised loss function like mean square error(MSE)calculates the distance between the predicted value and the original input. However, it is difficult to guarantee the effectiveness of the features only by optimizing there construction error.In order to make the learned features more effective for classification tasks, we optimize the contrastive loss function to make the features fromdifferent views of the same sample consistent. This makes the features of the same classaggregate with each other, and the features of different classes are far away from eachother. Therefore, the features obtained by optimizing the contrastive loss function ofdifferent views could effectively improve the classification accuracy. We use a deepCNN as the base feature extractor.We call this proposed method deep multiview learning. Therefore, the proposed method belongs to the category of unsupervised learning, which could alleviate the lack of labeled training samples. Finally, a conventional machine learning method(e.g.,supportvectormachine)is used to complete the classification task in the learned latent space. To demonstrate the effectiveness of theproposed method, extensive experiments are carried on four widely used hyperspectraldata sets. The experimental results demonstrate that the proposed method could improvethe classificationaccuracywith smallsamples.