Recently, Transformer-based methods have achieved significant improvements over convolutional neural networks (CNNs) in single image deraining, due to the powerful ability of modeling non-local information. In fact, r...
详细信息
ISBN:
(纸本)9798350338935
Recently, Transformer-based methods have achieved significant improvements over convolutional neural networks (CNNs) in single image deraining, due to the powerful ability of modeling non-local information. In fact, rich local-global information representations are equally important for better satisfying rain removal. In this paper, we propose an effective image deraining method by integrating a CNN model into the Transformer backbone to accelerate network convergence, called Multi-scale Dilated-convolution Transformer (MDT), which fully leverages the learning capabilities of Transformers on non-local features, seamlessly integrating local detail extraction and global structural representation. The fundamental building unit of our framework is the Multi-scale Dilated-convolution Transformer Block (MDTB) with different dilation rates, which consists of the Dilconv Self-Attention (DSA) and the Dilconv Feed-Forward Network (DFN). Specifically, the former processes the contextual information via dilated convolutions and enables the model to emphasize spatially-varying rain distribution features, while the latter integrates the dual-branch information to facilitate the local feature learning for better feature aggregation. Extensive evaluations demonstrate that our model reaches superior performance, significantly improving the image deraining quality.
This paper presents a multiscale graph construction method using both graph and signal features. Multiscale graph is a hierarchical representation of the graph, where a node at each level indicates a cluster in a fine...
详细信息
ISBN:
(纸本)9798350372267;9798350372250
This paper presents a multiscale graph construction method using both graph and signal features. Multiscale graph is a hierarchical representation of the graph, where a node at each level indicates a cluster in a finer resolution. To obtain the hierarchical clusters, existing methods often use graph clustering;however, they may ignore signal variations. As a result, these methods could fail to detect the clusters having similar features on nodes. In this paper, we consider graph and node-wise features simultaneously for multiscale clustering of a graph. With given clusters of the graph, the clusters are merged hierarchically in three steps: 1) Feature vectors in the clusters are extracted. 2) Similarities among cluster features are calculated using optimal transport. 3) A variable k-nearest neighbor graph (VkNNG) is constructed and graph spectral clustering is applied to the VkNNG to obtain clusters at a coarser scale. Additionally, the multiscale graph in this paper has non-local characteristics: Nodes with similar features are merged even if they are spatially separated. In experiments on multiscale image and point cloud segmentation, we demonstrate the effectiveness of the proposed method.
In this paper, we propose a no-reference image quality assessment method based on non-local features learned by a graph neural network (GNN). The proposed quality assessment framework is rooted in the view that the hu...
详细信息
ISBN:
(数字)9781665471893
ISBN:
(纸本)9781665471893
In this paper, we propose a no-reference image quality assessment method based on non-local features learned by a graph neural network (GNN). The proposed quality assessment framework is rooted in the view that the human visual system perceives image quality with long-dependency constructed among different regions, inspiring us to explore the non-local interactions in quality prediction. Instead of relying on convolutional neural network (CNN) based quality assessment methods that primarily focus on local field features, the GNN aiming for non-local quality perception facilitates modeling such long-dependency. In particular, we first adopt superpixel segmentation for the graph nodes construction. Subsequently, a spatial attention module is proposed to integrate the long- and short-range dependencies among the nodes of the whole image. The learned non-local features are finally combined with the local features extracted by the pre-trained CNN, achieving superior performance to the features utilized individually. Experimental results on intra-dataset and cross-dataset settings verify our proposed method's effectiveness and advanced generalization capability. Source codes are publicly accessible at https://***/SuperBruceJia/NLNet-IQA for scientific reproducible research.
Using MAGPIE (Machine Automated General Performance Improvement via Evolution of software), we measure the impact of genetic improvement (GI) on a non-deterministic deeply nested PARSEC VIPS parallel computing multi-t...
详细信息
ISBN:
(纸本)9798400705731
Using MAGPIE (Machine Automated General Performance Improvement via Evolution of software), we measure the impact of genetic improvement (GI) on a non-deterministic deeply nested PARSEC VIPS parallel computing multi-threaded imageprocessing benchmark written in C. More than 53% of mutants compile and generate identical results to the original program. We find about 10% Failed Disruption Propagation (FDP). Excluding internal errors and asserts, almost all changes deeper than 30 nested functions which are Executed and Infect data or change control are not Propagated to the output, i.e. these deep PIE changes have no external effect. Suggesting (where it relies on testing) automatic software engineering on deeply nested code will be hard.
Constructive approaches, principles, models for optimizing the identification of micro-objects based on the use of statistical, dynamic models with mechanisms for filtering noise, foreign particles on images of medica...
详细信息
The paper explores the development and application of linear operators for filtering digital images, particularly through the use of fifth-order polynomial splines. The authors emphasize the importance of high-quality...
详细信息
The paper explores the development and application of linear operators for filtering digital images, particularly through the use of fifth-order polynomial splines. The authors emphasize the importance of high-quality, high-speed imageprocessing in enhancing cybersecurity systems, which require accurate and real-time data analysis for applications such as military intelligence and UAV navigation. They propose using B-splines, which offer precise approximation capabilities while maintaining computational efficiency. The paper outlines a method for representing these splines in a manner that reduces computational complexity, making them suitable for real-time processing in environments with high data throughput demands. The research highlights the advantages of using wider window filters and linear operators to maintain image accuracy and fidelity. It demonstrates how these techniques can be effectively applied to a range of signal processing tasks beyond image filtering, including audio and telecommunications. The experimental results presented in the paper indicate that the proposed filters significantly improve the detection and accuracy of special points in digital images, which is crucial for UAV navigation and other critical applications. The authors conclude that the use of these advanced filtering techniques can lead to more efficient and sophisticated imageprocessing systems, with potential applications across various fields. In addition to the practical applications, the paper also discusses the theoretical underpinnings of splinebased filters, including the derivation of low- and high-frequency filter masks. The authors propose that these filters can be used to perform both sub-band filtering and multi-scale analysis of digital images, leading to improved data processing and analysis capabilities. The research contributes to the ongoing development of efficient imageprocessing techniques that are essential for modern technology and various real-time appl
In the problem of flexible job shop scheduling, conventional genetic algorithms exhibit limited exploration capabilities and are prone to getting stuck in local optima. Consequently, this study aims to enhance the tra...
详细信息
A new approach for locally-adaptive processing of color RGB images represented as tensors of size M × N × 3, is offered in this work. Unlike the famous similar methods of the kind, the processing here is exe...
详细信息
This paper intends to optimize the overall implementing process of federated learning (FL) in practical edge computing systems. First, we present a general FL algorithm, namely GenQSGD+, whose parameters include the n...
详细信息
ISBN:
(数字)9781665494557
ISBN:
(纸本)9781665494557
This paper intends to optimize the overall implementing process of federated learning (FL) in practical edge computing systems. First, we present a general FL algorithm, namely GenQSGD+, whose parameters include the numbers of global and local iterations, mini-batch size, and step size sequence. Then, we analyze the convergence of GenQSGD+ with arbitrary algorithm parameters. Next, we optimize all the algorithm parameters of GenQSGD+ to minimize the energy cost under the constraints on the time cost, convergence error, and step size sequence. The resulting optimization problem is challenging due to its non-convexity and the presence of a dimension-varying vector variable and non-differentiable constraint functions. We transform the complicated problem into a more tractable nonconvex problem using the structural properties of the original problem and propose an iterative algorithm using general inner approximation (GIA) and complementary geometric programming (CGP) to obtain a KKT point. Finally, we numerically demonstrate remarkable gains of optimization-based GenQSGD+ over typical FL algorithms and the advancement of the proposed optimization framework for federated edge learning.
Cell cultures suspended in bioreactors in a fluid environment are the basis for cell expansion and important medical products manufacturing. Assessing local cell distribution within bioreactors may provide information...
详细信息
暂无评论