Explore the vast range of titles available.

Palmprint is deemed as one of the most important biometric modalities. Many coding-based palmprint recognition methods have achieved satisfactory recognition performance, which can be free from training and require low storage cost and computational complexity. Downsampling is typically used to improve real-time ability, reduce the storage cost and improve the discriminative ability. Unfortunately, downsampling was not fully considered and studied. In this paper, we propose downsampling in uniformly-spaced windows (DUSW) and conduct it on two state-of-the-art downsampling methods as their reformative versions, dubbed uniformly-spaced extreme downsampling method (U-EDM) and uniformly-spaced democratic voting downsampling method (U-DVDM). In DUSW, the upper-left four pixels rather than all pixels in each block are selected to jointly decide the winner whose value is used as the representative feature of this block. DUSW overcomes the dictatorship of a single pixel and simultaneously ensures the sufficient spatial distance between the adjacent winners. Thus, DUSW reduces the correlation between the adjacent winners, and accordingly improves the discrimination and robustness. Meanwhile, the computational complexity is only 1/4 of the original downsampling methods for the representative reduction. The sufficient experiments demonstrate that DUSW can be easily embedded into the existing downsampling methods of coding-based palmprint recognition and improve their recognition performances.

In response to the challenge of handling large-scale 3D point cloud data, downsampling is a common approach, yet it often leads to the problem of feature loss. We present a dynamic downsampling algorithm for 3D point cloud maps based on an improved voxel filtering approach. The algorithm consists of two modules, namely, dynamic downsampling and point cloud edge extraction. The former adapts voxel downsampling according to the features of the point cloud, while the latter preserves edge information within the 3D point cloud map. Comparative experiments with voxel downsampling, grid downsampling, clustering-based downsampling, random downsampling, uniform downsampling, and farthest-point downsampling were conducted. The proposed algorithm exhibited favorable downsampling simplification results, with a processing time of 0.01289 s and a simplification rate of 91.89%. Additionally, it demonstrated faster downsampling speed and showcased improved overall performance. This enhancement not only benefits productivity but also highlights the system’s efficiency and effectiveness.

In recent years, numerous CNN-based light field (LF) image super-resolution (SR) methods have been developed. However, due to the downsampling inconsistency between low-resolution (LR) testing LF images and LR training LF images, they may suffer from quality degradation. To address this quality degradation issue, this paper proposes a downsampling consistency correction-based (DCC-based) quality enhancement method. Firstly, a quality-based voting strategy is introduced to identify the downsampling scheme used in the training step. Next, a cascaded Swin Transformer-based recognizer is proposed to identify the downsampled position and downsampling scheme used in the LR testing LF image, after which the proposed DCC-based method is employed to significantly improve the quality of the upsampled LF image. Comprehensive experiments have been conducted on typical LF image datasets to demonstrate the significant quality improvement achieved by our method in comparison to state-of-the-art LF SR methods.

Almost all computer vision tasks rely on convolutional neural networks and transformers, both of which require extensive computations. With the increasingly large size of images, it becomes challenging to input these images directly. Therefore, in typical cases, we downsample the images to a reasonable size before proceeding with subsequent tasks. However, the downsampling process inevitably discards some fine-grained information, leading to network performance degradation. Existing methods, such as strided convolution and various pooling techniques, struggle to address this issue effectively. To overcome this limitation, we propose a generalized downsampling module, Adaptive Separation Fusion Downsampling (ASFD). ASFD adaptively captures intra- and inter-region attentional relationships and preserves feature representations lost during downsampling through fusion. We validate ASFD on representative computer vision tasks, including object detection and image classification. Specifically, we incorporated ASFD into the YOLOv7 object detection model and several classification models. Experiments demonstrate that the modified YOLOv7 architecture surpasses state-of-the-art models in object detection, particularly excelling in small object detection. Additionally, our method outperforms commonly used downsampling techniques in classification tasks. Furthermore, ASFD functions as a plug-and-play module compatible with various network architectures.

Vision-based environmental perception has demonstrated significant promise for autonomous driving applications. However, the traditional unidirectional feature flow in many perception networks often leads to inadequate information propagation, which hinders the system’s ability to comprehensively perceive complex driving environments. Issues such as similar objects, illumination variations, and scale differences aggravate this limitation, introducing noise and reducing the reliability of the perception system. To address these challenges, we propose a novel Attention-Aware Upsampling-Downsampling Network (AUDNet). AUDNet utilizes a bidirectional feature fusion structure, incorporating a multi-scale attention upsampling module (MAU) to enhance the fine details in high-level features by guiding the selection of feature information. Additionally, the multi-scale attention downsampling module (MAD) is designed to reinforce the semantic understanding of low-level features by emphasizing relevant spatial dfigureetails. Extensive experiments on a large-scale, real-world driving dataset demonstrate the superior performance of AUDNet, particularly in multi-task environment perception in complex and dynamic driving scenarios.

Bearing faults in rotating machinery can lead to significant economic losses due to downtime and pose serious safety risks. Accurate fault diagnosis is crucial for effective condition monitoring. Traditional methods for diagnosing bearing faults under noisy conditions often rely on complex data preprocessing and struggle to maintain accuracy in high-noise environments. To address this challenge, this paper proposes an end-to-end Discrete Wavelet Integrated Convolutional Residual Neural Network (DWCResNet) for bearing fault diagnosis. The model incorporates Discrete Wavelet Transform (DWT) layers to replace traditional downsampling operations in convolutional neural networks, decomposing input signals into low-frequency and high-frequency components to effectively remove high-frequency noise and extract fault features, thereby improving diagnostic performance. The cyclic learning rate strategy enhances training efficiency. Experiments conducted on the Case Western Reserve University (CWRU) and Paderborn University (PU) bearing datasets demonstrate that DWCResNet achieves higher diagnostic accuracy and noise robustness under various conditions, providing an efficient solution for bearing fault diagnosis in complex noisy environments.

Intelligent vehicles should not only be able to detect various obstacles, but also identify their categories so as to take an appropriate protection and intervention. However, the scenarios of object detection are usually complex and changeable, so how to balance the relationship between accuracy and speed is a difficult task of object detection. This paper proposes a multi-object detection algorithm using DarkNet-53 and dense convolution network (DenseNet) to further ensure maximum information flow between layers. Three 8-layer dense blocks are used to replace the last three downsampling layers in DarkNet-53 structure, so that the network can make full use of multi-layer convolution features before prediction. The loss function of coordinate prediction error in YOLOv3 is further improved to improve the detection accuracy. Extensive experiments are conducted on the public KITTI and Pascal VOC datasets, and the results demonstrate that the proposed algorithm has better robustness, and the network model is more suitable for the traffic scene in the real driving environment and has better adaptability to the objects with long distance, small size and partial occlusion.

Quantum machine learning is a promising application of quantum computing for data classification. However, most of the previous research focused on binary classification, and there are few studies on multi-classification. The major challenge comes from the limitations of near-term quantum devices on the number of qubits and the size of quantum circuits. In this paper, we propose a hybrid quantum neural network to implement multi-classification of a real-world dataset. We use an average pooling downsampling strategy to reduce the dimensionality of samples, and we design a ladder-like parameterized quantum circuit to disentangle the input states. Besides this, we adopt an all-qubit multi-observable measurement strategy to capture sufficient hidden information from the quantum system. The experimental results show that our algorithm outperforms the classical neural network and performs especially well on different multi-class datasets, which provides some enlightenment for the application of quantum computing to real-world data on near-term quantum processors.

The integration of Multi-layer Perceptrons (MLPs), Convolutional Neural Networks (CNNs), and attention mechanisms has been demonstrated to significantly enhance model performance across various computer vision tasks. In this paper, a novel lightweight neural network architecture, FasterMLP, is proposed to achieve high computational efficiency and accuracy, particularly in resource-constrained and real-time applications. FasterMLP is designed to combine the local connectivity and weight-sharing properties of CNNs with the global feature representation capabilities of MLPs, while feature extraction is enhanced through the Convolutional Block Attention Module and spatial dimensions are effectively reduced using Haar wavelet downsampling without sacrificing critical feature information. The architecture, structured into four stages, has been rigorously evaluated on multiple benchmarks. On the ImageNet-1K dataset, a top-1 accuracy 3.9% higher than that of MobileViT-XXS is achieved by FasterMLP-S, while being 2 and 2.7 faster on GPU and CPU, respectively. On the COCO dataset, the performance of FasterMLP-L is shown to be comparable to FasterNet-L with significantly fewer parameters, and on the Cityscapes dataset, a mean Intersection-over-Union of 81.7% is achieved, surpassing existing methods such as CCNet and DANet. These results demonstrate that FasterMLP can effectively balance computational efficiency and accuracy, making it particularly suitable for visual perception tasks in resource-constrained and real-time environments such as autonomous driving. Code is available at https://github.com/windisl/FasterMLP .

In view of the difficulty of using raw 3D point clouds for component detection in the railway field, this paper designs a point cloud segmentation model based on deep learning together with a point cloud preprocessing mechanism. First, a special preprocessing algorithm is designed to resolve the problems of noise points, acquisition errors, and large data volume in the actual point cloud model of the bolt. The algorithm uses the point cloud adaptive weighted guided filtering for noise smoothing according to the noise characteristics. Then retaining the key points of the point cloud, this algorithm uses the octree to partition the point cloud and carries out iterative farthest point sampling in each partition for obtaining the standard point cloud model. The standard point cloud model is then subjected to hierarchical multi-scale feature extraction to obtain global features, which are combined with local features through a self-attention mechanism, while linear interpolation is used to further expand the perceptual field of local features of the model as a basis for segmentation, and finally the segmentation is completed. Experiments show that the proposed algorithm could deal with the scattered bolt point cloud well, realize the segmentation of train bolt and background, and could achieve high segmentation accuracy, which has important practical significance for train safety detection.