This paper introduces a new approach for video denoising. Based on the idea of patch based low rank matrix completion, we improve the method by modeling noises with Mixture of Gaussians (MoG). By utilizing a series of different gaussian distributions to fit the representation of video noises without any assumptions on the statistical properties, the parameters of MoG are learned from video data automatically. It can deal with the fact that for most of the time, the real distribution of noises appeared in videos are unknown so that traditional methods do not work well without any priori knowledge. After the model and algorithm statements, we provide a group of experiments on real videos for comparisons with the state-of-art video denoising algorithm, which demonstrates the effectiveness and advantage of our approach.

Image clustering is a complex procedure that is significantly affected by the choice of the image representation. Generally speaking, image representations are generated by using handcraft features or trained neural networks. When dealing with high dimension data, these two representation methods cause two problems: i) the representation ability of the manually designed features is limited; ii) the non-representative and meaningless feature of a trained deep network may hurt the clustering performance. To overcome these problems, we propose a new clustering method which efficiently builds an image representation and precisely discovers the cluster assignments. Our main tools are an unsupervised representation learning method based on Deep Mutual Information Maximization (DMIM) system, and a clustering method based on self-training algorithm. Specifically speaking, to extract the informative representation of image data, we derive the maximum mutual information theory and propose a system to learn the maximum mutual information between the input images and the latent representations. To discover the clusters and assign each image a clustering label, a self-training mechanism is applied to cluster the learned representations. The superiority and validity of our algorithm are verified in a series of real-world image clustering experiments.

Because of the limitations of matrix factorization, such as losing spatial structure information, the concept of low-rank tensor factorization (LRTF) has been applied for the recovery of a low dimensional subspace from high dimensional visual data. However, existing methods often fail to tackle the real data which are corrupted by the noise with unknown distribution. In this paper, we propose a novel noise model to the tensor case for the LRTF task to overcome the drawbacks of existing models. This procedure treats the target data as high-order tensor directly and models the noise by a Mixture of Gaussians and a Markov Random Field, which is called MoG WLRTF MRF. The parameters in the model are estimated under the variational EM framework. Extensive experiments demonstrate the effectiveness of our method compared with other competing methods.