Clustering complex-shaped clusters is still chal lenging for most existing clustering algorithms. Herein, the peak-padding clustering algorithm (PeakPad)-clustering by padding density peaks with the minimum padding cost-is proposed. PeakPad executes clustering on the density surface and views complex-shaped clusters as combinations of highly associated single-peak clusters. The minimum padding cost that fully considers the surrounding context of a density peak is proposed to reflect a density peak's center potential, enabling PeakPad to have robust center detection performance. Unlike mean-shift (MSC), which detects centers based on their attributes in a complex-shaped density surface embedded in the high-dimensional space of density and features, PeakPad detects centers in a standard-shaped surface embedded in the 2-D density-change (DC) density space (composed of density and DC feature). Such standardization allows PeakPad to have fast and robust cluster center detection performance on complex-shaped clusters based on the minimum padding cost. Besides, PeakPad can provide a reasonable evaluation of the association between single-peak clusters by using the minimum padding cost. As a result, PeakPad can fast capture complex-shaped clusters, achieve robust center detection performance, and be suitable for large datasets. Benchmark test results on both synthetic and real datasets demonstrate the effectiveness of PeakPad.

In the magnetic actuation system, MAC (Magnetically Actuated Capsule) motion is disturbed by gastrointestinal resistance, and dynamic constraints exist between MAC and EPM (External Permanent Magnet), leading to control difficulties. This paper presents a data-driven control method for magnetic actuation system. Firstly, a data-driven modeling approach is proposed for addressing the modeling challenges, relying solely on MAC visual positioning input and EPM position output data. Secondly, To effectively track rapidly changing MAC desired trajectories, determining the upper bound of system inputs through analysis of the magnetic field relationship, and integrating it with adaptive parameter reset conditions, enables the establishment of dynamic constraints for the MAC system. Finally, dynamic compensation is applied to account for non-linear resistance terms and inaccuracies in data model representation. A dual-visual positioning experimental platform simulating the gastrointestinal environment is established to validate the proposed algorithm’s effectiveness in MAC trajectory tracking under different conditions.Note to Practitioners—This article aims to design a data-driven method that incorporates dynamic constraint relationships between MAC and EPM, enabling capsules to move rapidly within the gastrointestinal system. This paper aims to increase control speed, enabling the energy to be primarily used for capturing the lesion area rather than during motion, and improving MAC tracking for desired trajectory accuracy. To validate the algorithm’s effectiveness under various resistances and driving forces, a dual-camera positioning system is designed and tested on a simulated gastrointestinal platform. In future work, we plan to replace the external camera with an internal capsule camera for MAC localization and design a medical system that actively utilizes patient lesion data to drive WCE for examinations.

Graph clustering technique is highly effective in detecting complex-shaped clusters, in which graph building is a crucial step. Nevertheless, building a reasonable graph that can exhibit high connectivity within clusters and low connectivity across clusters is challenging. Herein, we design a max-ascent-angle graph called the “Y-graph”, a high-sparse graph that automatically allocates dense edges within clusters and sparse edges across clusters, regardless of their shapes or dimensionality. In the graph, every point$x$is allowed to connect its nearest higher-density neighbor$\delta$, and another higher-density neighbor$\gamma$, satisfying that the angle$\angle \delta x\gamma$is the largest, called “max-ascent-angle”. By seeking the max-ascent-angle, points are automatically connected as the Y-graph, which is a reasonable graph that can effectively balance inter-cluster connectivity and intra-cluster non-connectivity. Besides, an edge weight function is designed to capture the similarity of the neighbor probability distribution, which effectively represents the density connectivity between points. By employing the Normalized-Cut (Ncut) technique, a Ncut-Y algorithm is proposed. Benefiting from the excellent performance of Y-graph, Ncut-Y can fast seek and cut the edges located in the low-density boundaries between clusters, thereby, capturing clusters effectively. Experimental results on both synthetic and real datasets demonstrate the effectiveness of Y-graph and Ncut-Y.

Since most existing single-prototype clustering algorithms are unsuitable for complex-shaped clusters, many multi-prototype clustering algorithms have been proposed. Nevertheless, the automatic estimation of the number of clusters and the detection of complex shapes are still challenging, and to solve such problems usually relies on user-specified parameters and may be prohibitively time-consuming. Herein, a stable-membership-based auto-tuning multi-peak clustering algorithm (SMMP) is proposed, which can achieve fast, automatic, and effective multi-prototype clustering without iteration. A dynamic association-transfer method is designed to learn the representativeness of points to sub-cluster centers during the generation of sub-clusters by applying the density peak clustering technique. According to the learned representativeness, a border-link-based connectivity measure is used to achieve high-fidelity similarity evaluation of sub-clusters. Meanwhile, based on the assumption that a reasonable clustering should have a relatively stable membership state upon the change of clustering thresholds, SMMP can automatically identify the number of sub-clusters and clusters, respectively. Also, SMMP is designed for large datasets. Experimental results on both synthetic and real datasets demonstrated the effectiveness of SMMP.

Most existing clustering algorithms require presetting cluster number and often fail to capture complex shapes. Herein, we propose a clustering algorithm by pruning a density-boosting cluster tree of density mounts—DEnsity MOuntains Separation clustering algorithm (DEMOS). A cluster is assumed to be a density-connected area with multiple (or a single) density mounts (i.e., single-peak clusters) and a relatively large dis-connectivity from density-connected areas of higher densities. Based on this assumption, DEMOS can easily detect the number of clusters and robustly reconstruct their complex shapes. It first builds the dataset into a peak graph, where each density peak represents a density mount. A multi-valley-link-based connectivity estimation method is embedded to efficiently estimate the connectivity between density peaks during peak graph building. Then, by applying a new linkage metric designed based on our assumption, DEMOS builds density mounts into a reasonably density-boosting cluster tree. After obtaining a robust center detection in a clarity-enhancing decision graph (i.e., a two-dimensional plot for detecting centers), DEMOS prunes the cluster tree into final clusters to finish clustering. Experimental results on both synthetic and real datasets demonstrated the effectiveness of DEMOS and its applicability to large-scale data clustering.

Abstract Automatic classification of diseases in endoscopic images is essential to the improvement of diagnostic performance and the reduction of colorectal cancer mortality. However, due to the ambiguous boundary between background and foreground, abnormal classification in endoscopic images is still challenging. To tackle such a situation, an adaptive aggregation with self‐attention network (AASAN), including a global branch, a local branch, and a fusion branch, is proposed imitating the diagnosis process of endoscopists. On this basis, the self‐attention with relative position encoding (SA‐RPE) module is designed to capture long‐range dependencies and gather lesion neighborhood information. Furthermore, an adaptive aggregation feature (AAF) module is proposed and embedded into the fusion branch for final image label prediction, which is helpful to capture more discriminant features. Extensive experiments show that the classification accuracy of the authors' method on Kvasir public dataset reaches 96.37% in a fivefold cross‐validation, higher than the state‐of‐the‐art deep learning algorithms.

Fuzzy c-means (FCM) algorithm as a traditional clustering algorithm for image segmentation cannot effectively preserve local spatial information of pixels, which leads to poor segmentation results with inconsistent regions. For the remedy, superpixel technologies are applied, but spatial information preservation highly relies on the quality of superpixels. Density peak clustering algorithm (DPC) can reconstruct spatial information of arbitrary-shaped clusters, but its high time complexity$O(n^2)$and unrobust allocation strategy decrease its applicability for image segmentation. Herein, a fast density peak clustering method (PGDPC) based on the kNN distance matrix of data with time complexity$O(nlog(n))$is proposed. By using the peak-graph-based allocation strategy, PGDPC is more robust in the reconstruction of spatial information of various complex-shaped clusters, so it can rapidly and accurately segment images into high-consistent segmentation regions. Experiments on synthetic datasets, real and Wireless Capsule Endoscopy (WCE) images demonstrate that PGDPC as a fast and robust clustering algorithm is applicable to image segmentation.

This paper deals with the design of a sensing matrix along with a sparse recovery algorithm by utilizing the probability-based prior information for compressed sensing systems. With the knowledge of the probability for each atom of the dictionary being used, a diagonal weighted matrix is obtained and then the sensing matrix is designed by minimizing a weighted function such that the Gram of the equivalent dictionary is as close to the Gram of dictionary as possible. An analytical solution for the corresponding sensing matrix is derived that requires low computational complexity. We also exploit this prior information through the sparse recovery stage and propose a probability-driven orthogonal matching pursuit algorithm that improves the accuracy of the recovery. Simulations for synthetic data and application scenarios of video streaming are carried out to compare the performance of the proposed methods with some existing algorithms. The results reveal that the proposed compressed sensing (CS) approach outperforms existing CS systems.

This study deals with the issue of designing the sensing matrix for a compressed sensing (CS) system assuming that the dictionary is given. Traditionally, the measurement of small mutual coherence is considered to design the optimal sensing matrix so that the Gram of the equivalent dictionary is as close to the target Gram as possible, where the equivalent dictionary is not normalised. In other words, these algorithms are designed to solve the CS problem using an optimisation stage followed by normalisation. To achieve a global solution, a novel strategy of the sensing matrix design is proposed by using a gradient‐based method, in which the measure of real mutual coherence for the equivalent dictionary is considered. According to this approach, a minimised objective function based on alternating minimisation is also developed through searching the target Gram within a set of relaxed equiangular tight frames. Some experiments are done to compare the performance of the newly designed sensing matrix with the existing ones under the condition that the dictionary is fixed. For the simulations of synthetic data and real image, the proposed approach provides better signal reconstruction accuracy.

In this paper, a distributed compressed estimation (DCE) scheme is presented based on a distributed recursive-least squares algorithm for sparse signals and systems along with a sensing matrix design procedure based on compressive sensing techniques. The D-CE scheme consists of compression and decompression modules inspired by compressive sensing to perform distributed compressed estimation. A design procedure is developed under the DCE framework and a novel algorithm is developed to optimize the sensing matrix, which can further improve the performance of the proposed DCE and distributed adaptive algorithms. Simulations for a wireless sensor network show the advantages of the proposed scheme and algorithm in terms of convergence rate and mean square error performance.

This paper deals with the sensing matrix optimization problem for compressed sensing (CS) systems. Traditionally, the optimal sensing matrix is designed such that the Gram of the equivalent dictionary defined as the product of the sensing matrix and the dictionary is as close to a target Gram with some proper properties as possible. In this study, the sensing matrix is designed to make the equivalent dictionary approximate to a certain target frame. In addition, to avoid the sparse representation error (SRE) to be amplified in the measurement domain, a penalty term related to the SRE is included in the design criterion. An alternating minimization algorithm is proposed to solve the optimum sensing matrix problem, where the target frame is taken as the relaxed equiangular tight frame, which is constructed with a new method with the purpose of reducing the mutual coherence and maintaining the tightness of the frame, then the solution of the optimal sensing matrix is derived analytically with the target frame fixed. Experiments are carried out with synthetic data and real images, which demonstrate promising performance of the proposed algorithms and superiority of the CS system designed with the optimized sensing matrix to existing ones in terms of signal reconstruction accuracy.

This paper deals with designing sensing matrix for compressive sensing systems. Traditionally, the optimal sensing matrix is designed so that the Gram of the equivalent dictionary is as close as possible to a target Gram with small mutual coherence. A novel design strategy is proposed, in which, unlike the traditional approaches, the measure considers of mutual coherence behavior of the equivalent dictionary as well as sparse representation errors of the signals. The optimal sensing matrix is defined as the one that minimizes this measure and hence is expected to be more robust against sparse representation errors. A closed-form solution is derived for the optimal sensing matrix with a given target Gram. An alternating minimization-based algorithm is also proposed for addressing the same problem with the target Gram searched within a set of relaxed equiangular tight frame Grams. The experiments are carried out and the results show that the sensing matrix obtained using the proposed approach outperforms those existing ones using a fixed dictionary in terms of signal reconstruction accuracy for synthetic data and peak signal-to-noise ratio for real images.

In this work, we propose adaptive frequency-domain biased estimation algorithms with mechanisms to automatically adjust the shrink age factors. The proposed estimation algorithms improve the performance of the conventional least squares (LS) estimator in terms of mean-squared error (MSE), while requiring a very modest in crease in complexity. An extension of the Cramer-Rao Lower Bound (CRLB) is computed in order to serve as a performance benchmark for the MSE performance of the biased estimators. We consider an application of the proposed algorithms to single-carrier frequency domain equalization (SC-FDE) of direct-sequence ultra-wideband (DS-UWB) systems, in which the channel estimation is performed by the proposed algorithms. The simulation results show that the proposed algorithms significantly outperform existing methods.

A novel linear blind reduced-rank receiver based on the Joint Iterative Optimization (JIO) and the constrained constant modulus (CCM) design criterion is proposed for interference suppression in direct-sequence ultra-wideband (DS-UWB) systems. The proposed receiver consists of a projection matrix that performs dimensionality reduction and a reduced-rank filter that produces the output. The columns of the projection matrix and the reduced-rank filter are updated jointly at each time instant or iteration to minimize the CM cost function subject to a constraint. Recursive least-squares (RLS) algorithms are developed for the adaptive implementation. Simulation results show that the proposed scheme has excellent performance in suppressing the inter-symbol interference (ISI) and multiple access interference (MAI) with a low complexity.

This work presents joint iterative power allocation and interference suppression algorithms for DS-CDMA networks which employ multiple relays and the amplify and forward cooperation strategy. We propose a joint constrained optimization framework that considers the allocation of power levels across the relays subject to individual and global power constraints and the design of linear receivers for interference suppression. We derive constrained minimum mean-squared error (MMSE) expressions for the parameter vectors that determine the optimal power levels across the relays and the parameters of the linear receivers. In order to solve the proposed optimization problems efficiently, we develop recursive least squares (RLS) algorithms for adaptive joint iterative power allocation, and receiver and channel parameter estimation. Simulation results show that the proposed algorithms obtain significant gains in performance and capacity over existing schemes.

In this work, we propose a conjugate gradient (CG) based structured channel estimation (SCE) scheme for single-carrier frequency domain equalization (SC-FDE) in multiuser direct-sequence ultra-wideband (DS-UWB) systems. The minimum mean square error (MMSE) linear detection strategy is used and a cyclic prefix is employed. We perform the adaptive channel estimation in the frequency domain and implement the despreading in the time domain after the FDE. In this scheme, the linear MMSE detection requires the knowledge of the number of users and the noise variance. For this purpose, we propose algorithms for estimating these parameters. A CG adaptive algorithm is then developed for the SCE scheme. With lower complexity than the recursive least squares (RLS) algorithm and better performance than the Least mean squares (LMS) algorithm, the SCE-CG achieves a better tradeoff between the complexity and the performance.

We consider a two-stage framework for linear interference mitigation, in which a transformation performs dimensionality reduction followed by a reduced-rank filter. A generic reduced-rank scheme that jointly optimizes the transformation and the reduced-rank filter by using the minimum mean squared error (MMSE) criterion is investigated. Then, we impose constraints on the design of the transformation and propose the switched approximations of adaptive basis functions (SAABF) scheme, in which the transformation is chosen instantaneously from a set of mapping matrices and adaptive basis functions. Least-mean squares (LMS) algorithms and model-order selection algorithms are also proposed. A complexity analysis shows that the proposed scheme is significantly simpler than the existing reduced-rank schemes. Simulations show remarkable interference mitigation performance in direct-sequence ultra-wideband (DS-UWB) systems.

Here, the authors propose two adaptive detection schemes based on single-carrier frequency-domain equalisation (SC‐FDE) for multiuser direct-sequence ultra-wideband systems, which are termed structured channel estimation (SCE) and direct adaptation (DA). Both schemes use the minimum mean square error (MMSE) linear detection strategy and employ a cyclic prefix. In the SCE scheme, adaptive channel estimation is performed in the frequency domain and the despreading is implemented in the time domain after the FDE. In this scheme, the MMSE detection requires the knowledge of the number of users and the noise variance. For this purpose, simple algorithms are proposed for estimating these parameters. In the DA scheme, the interference suppression task is fulfilled with only one adaptive filter in the frequency domain and a new signal expression is adopted to simplify the design of such a filter. Least mean squares, recursive least squares and conjugate gradient adaptive algorithms are then developed for both schemes. A complexity analysis compares the computational complexity of the proposed algorithms and schemes, and simulation results for the downlink illustrate their performance.