High quality medical images are not only an important basis for doctors to carry out clinical diagnosis and treatment, but also conducive downstream tasks such as image analysis. Although many enhancement methods have achieved good results, some of them still shortcomings in homogenizing illumination distribution maintaining texture details, even introduce boundary artifact noise. In order deal with these problems, this paper proposes a multi-scale attention generative adversarial network...

The ensemble of fuzzy clustering can address the problems presented in base clustering, such as fluctuations results due to random initialization and performance degradation outliers. However, ensembles is still hampered by some challenges that include misaligned membership matrices, loss information cosimilarity matrix, large storage space, unstable an additional reclustering, need for original data assistance, etc. To these issues, we propose a parameter-free robust framework clustering....

Uncertain data clustering has been recognized as an essential task in the research of mining. Many centralized algorithms are extended by defining new distance or similarity measurements to tackle this issue. With fast development network applications, these methods show their limitations conducting a large dynamic distributed peer-to-peer due privacy and security concerns technical constraints brought distributive environments. In paper, we propose novel uncertain algorithm, which global...

Categorical data are widely available in many real-world applications, and to discover valuable patterns such by clustering is of great importance. However, the lack a decent quantitative relationship among categorical values makes traditional approaches, which usually developed for numerical data, perform poorly on datasets. To solve this problem boost performance we propose novel fuzzy model article. At first, approximating maximum posteriori (MAP) estimation discrete distribution...

The traditional collaborative fuzzy clustering can effectively perform data in distributed peer-to-peer networks, which is an impossible task to complete for the centralized methods due privacy and security requirements or network transmission technology constraints. But it will increase number of iterations lead lower efficiency clustering. Moreover, mechanism hidden iterative process cannot be well revealed explained. In this article, a novel series transfer algorithms are proposed solve...

Tensor-based robust principal component analysis (PCA) methods are efficient to discover the low-rank part of a hyperspectral image for reducing redundant information and guarantee good classification results. However, current cannot remove noise adequately, residual remaining in limits further improvement performance. Thus, enhancing robustness is important helpful tensor-based PCA (RPCA) process images. To this end, we propose RPCA method with locality preserving graph frontal slice...

The graph-information-based fuzzy clustering has shown promising results in various datasets. However, its performance is hindered when dealing with high-dimensional data due to challenges related redundant information and sensitivity the similarity matrix design. To address these limitations, this article proposes an implicit k-means (FKMs) model that enhances graph-based for data. Instead of explicitly designing a matrix, our approach leverages partition result obtained from FKMs generate...

Kernel clustering has the ability to get inherent nonlinear structure of data. But high computational complexity and unknown representation kernel space make it unavailable for data in distributed peer-to-peer (P2P) networks. To solve this issue, we propose a new series random feature-based collaborative algorithms article. In most basic algorithm, each node P2P network first maps its into low-dimensional feature with approximation given by using Fourier mapping method. Then, independently...

Contrastive-based clustering models usually rely on a large number of negative pairs to capture uniform representations, which requires batch size and high computational complexity. In contrast, some self-supervised methods perform non-contrastive learning discriminative representations only with positive pairs, but suffer from the collapse clustering. To solve these issues, novel end-to-end model is proposed in this paper. The basic network first modified, followed by incorporation Softmax...

Deep multiview clustering provides an efficient way to analyze the data consisting of multiple modalities and features. Recently, autoencoder (AE)-based deep algorithms have attracted intensive attention by virtue their rewarding capabilities extracting inherent Nevertheless, most existing methods are still confronted several problems. First, usually contains abundant cross-view information, thus parallel performing individual AE for each view directly combining extracted latent together can...

Classifying antimicrobial peptides(AMPs) from the vast array of peptides mined metagenomic sequencing data is a significant approach to addressing issue antibiotic resistance. However, current AMP classification methods, primarily relying on sequence-based data, neglect spatial structure peptides, thereby limiting accurate AMPs. Additionally, number known AMPs significantly lower than that non-AMPs, leading imbalanced datasets reduce predictive accuracy for To alleviate these two...

Uncertain data clustering is one significant research in mining. Many similarity measurements of uncertain objects are proposed. Traditional methods can be extended with these new measurements. In this paper, we propose a fuzzy c-medoids method for clustering, named UFC-medoids. The JS-divergence used as the measurement between algorithm. experiments on synthetic datasets, presented algorithm has shown good performance.

The traditional clustering algorithms can not effectively deal with the when data for current task are enough. In this paper, we utilize transfer learning to assist entropy-weighted fuzzy c-means clustering. centers and corresponding weights of dimensions learned from known domain used in new objective function unknown Experiments on synthetic sets have demonstrated superiority algorithm.

Traditional kernel-based fuzzy clustering helps to discover the complex structures hidden in data, but it suffers from high computational complexity. This paper introduces random Maclaurin feature (RMF) into clustering, and proposes a feature-based algorithm (RMFCM). In this method, RMF is used approximate polynomial kernel, with linear complexity conducted generated space. The experiments carried out on four synthetic datasets UCI real-world prove effectiveness efficiency of proposed RMFCM...