End-to-end unsupervised clustering neural networks for image clustering

Abstract

<p>In this paper, we propose a new clustering module that can be trained jointly with existing neural network layers. Specifically, we have designed a generic clustering module with a competitive update mechanism. The module consists of a Gaussian unit and a maximum pooling layer. The Gaussian unit forward propagation conforms to the joint Gaussian distribution and contains two sets of trainable parameters. It requires no tedious setup and has a plug-and-play feature. To improve the representation capability of the network, we used an auto-encoder to extract the hidden semantics of the input features and combined the clustering module with the auto-encoder to construct an end-to-end unsupervised clustering neural network. We refer to this as HGL_CAE(High-dimensional Gaussian distribution layers combined with convolutional autoencoders).The network is highly adaptable to different input feature dimensions and can cope with situations where the number of clusters cannot be determined in advance. We conducted experiments on the MNIST dataset and the Fashion_MNIST dataset with a clustering accuracy of 93.38% and 72.83% respectively. It is highly competitive with existing methods. </p>

Determination of the number of clusters on images from space optic-electronic observation systems using the k-means algorithm

“…In [19], it is proposed to use an end-to-end unsupervised clustering neural network for image segmentation. The essence of [19] is the development of a separate clustering module consisting of a maximum unifying layer and a Gaussian block containing two sets of training parameters.…”

“…In [19], it is proposed to use an end-to-end unsupervised clustering neural network for image segmentation. The essence of [19] is the development of a separate clustering module consisting of a maximum unifying layer and a Gaussian block containing two sets of training parameters. The advantage of [19] is the possibility of simultaneous training of a separate clustering module with other layers of the neural network, adaptability to the situation when the number of clusters is not known in advance.…”

“…The essence of [19] is the development of a separate clustering module consisting of a maximum unifying layer and a Gaussian block containing two sets of training parameters. The advantage of [19] is the possibility of simultaneous training of a separate clustering module with other layers of the neural network, adaptability to the situation when the number of clusters is not known in advance. The disadvantage of [19] is the need to use an autoencoder in order to obtain the hidden semantics of input features.…”

“…The advantage of [19] is the possibility of simultaneous training of a separate clustering module with other layers of the neural network, adaptability to the situation when the number of clusters is not known in advance. The disadvantage of [19] is the need to use an autoencoder in order to obtain the hidden semantics of input features.…”