Displays最新文献

Breech face impressions are an essential type of physical evidence in forensic investigations. However, their surface morphology is complex and varies based on the machining method used on the gun’s breech face, making traditional handcrafted local feature-based methods exhibit high false rates and are unsuitable for striated impressions. We proposed a deep local feature-based method for firearm identification utilizing Detector-Free Local Feature Matching with Transformers (LoFTR). This method removes the module of feature point detection and directly utilizes self and cross-attention layers in the Transformer to transform the convolved coarse-level feature maps into a series of dense feature descriptors. Subsequently, matches with high confidence scores are filtered based on the score matrix calculated from the dense descriptors. Finally, the screened initial matches are refined into the convolved fine-level features, and a correlation-based approach is used to obtain the exact location of the match. Validation tests were conducted using three authoritative sets of the breech face impressions datasets provided by the National Institute of Standards and Technology (NIST). The validation results show that, compared with the traditional handcrafted local-feature based methods, the proposed method in this paper yields a lower identification error rate. Notably, the method can not only deal with granular impressions, but can also be applied to the striated impressions. The results indicate that the method proposed in this paper can be utilized for comparative analysis of breech face impressions, and provide a new automatic identification method for forensic investigations.

The real-time and accurate detection of three-dimensional (3D) objects based on LiDAR is a focal problem in the field of autonomous driving environment perception. Compared to two-stage and anchor-based 3D object detection methods that suffer from inference latency challenges, single-stage anchor-free 3D object detection approaches are more suitable for deployment in autonomous driving vehicles with the strict real-time requirement. However, they face the issue of insufficient spatial awareness, which can result in detection errors such as false positives and false negatives, thereby increasing the potential risks of autonomous driving. In response to this, we focus on enhancing the spatial awareness of CenterPoint, a widely used single-stage anchor-free 3D object detector in the industry. Considering the limited allocation of computational resources and the performance bottleneck caused by pillar encoder, we propose an efficient SSDCM backbone to strengthen feature representation and extraction. Furthermore, a simple BGC neck is devised to weight and exchange contextual information in order to deeply fuse multi-scale features. Combining improved backbone and neck networks, we construct a single-stage anchor-free 3D object detection model with spatial awareness enhancement, named CenterPoint-Spatial Awareness Enhancement (CenterPoint-SAE). We evaluate CenterPoint-SAE on two large-scale and challenging autonomous driving datasets, nuScenes and Waymo. It achieves 53.3% mAP and 62.5% NDS on nuScenes detection benchmark, and runs inference at a speed of 11.1 FPS. Compared to the baseline, the upgraded networks deliver a performance improvement of 1.6% mAP and 1.2% NDS at minor cost. Notably, on Waymo dataset, our method achieves competitive detection performance compared to two-stage and point-based methods.

Visually assistive systems for the visually impaired play a pivotal role in enhancing the quality of life for the visually impaired. Assistive technologies for the visually impaired have undergone a remarkable transformation with the advent of deep learning and sophisticated assistive devices. In particular, the paper utilizes the latest machine translation models and techniques to accomplish the Chinese–Braille translation task, providing convenience for visually impaired individuals. The Traditional end-to-end Chinese–Braille translation approach incorporates Braille dots and Braille word segmentation symbols as tokens within the model’s vocabulary. However, our findings reveal that Braille word segmentation is significantly more complex than Braille dot prediction. The paper proposes a novel Two-Part Loss (TPL) method that treats these tasks distinctly, leading to significant accuracy improvements. To enhance translation performance further, we introduce a BERT-Enhanced Segmentation Transformer (BEST) method. BEST leverages knowledge distillation techniques to transfer knowledge from a pre-trained BERT model to the translate model, mitigating its limitations in word segmentation. Additionally, soft label distillation is employed to improve overall efficacy further. The TPL approach achieves an average BLEU score improvement of 1.16 and 5.42 for Transformer and GPT models on four datasets, respectively. In addition, The work presents a two-stage deep learning-based translation approach that outperforms traditional multi-step and end-to-end methods. The proposed two-stage translation method achieves an average BLEU score improvement of 0.85 across four datasets.

Orthodontic treatment is a subject of prevention, treatment and prognostic care for malocclusion. The complexity and multimodality of orthodontic treatment restrict the development of intelligent orthodontics, due to the different growth and development characteristics of patients with different treatment stages and the different prognosis of treatment at different stages. Digital twin technology can effectively solve the problems of orthodontics ，due to its ability to interpret the deep information of big data. Building upon medical knowledge, this paper succinctly summarizes the application of digital twin technology in key areas of orthodontics, including precision Medicine, personalized orthodontic treatment, prediction of soft and hard tissues before and after orthodontic treatment, and the orthodontic cloud platform. The study provides a feasibility analysis of an intelligent predictive model for orthodontic treatment under multimodal fusion, offering a robust solution for establishing a comprehensive digital twin-assisted diagnostic paradigm.

Learning node representation on heterophilous graphs has been challenging due to nodes with diverse labels/attributes being connected. The main idea is to balance contributions between the center node and neighborhoods. However, existing methods failed to make full use of personalized contributions of different neighborhoods based on whether they own the same label as the center node, making it necessary to explore the distinctive contributions of similar/dissimilar neighborhoods. We reveal that both similar/dissimilar neighborhoods have positive impacts on feature aggregation under different homophily ratios. Especially, dissimilar neighborhoods play a significant role under low homophily ratios. Based on this, we propose LAAH, a label-aware aggregation approach for node representation learning on heterophilous graphs. LAAH separates each center node from its neighborhoods and generates their own node representations. Additionally, for each neighborhood, LAAH records its label information based on whether it belongs to the same class as the center node and then aggregates its effective feature in a weighted manner. Finally, a learnable parameter is used to balance the contributions of each center node and all its neighborhoods, leading to updated representations. Extensive experiments on 8 real-world heterophilous datasets and a synthetic dataset verify that LAAH can achieve competitive or superior accuracy in node classification with lower parameter scale and computational complexity compared with the SOTA methods. The code is released at GitHub: https://github.com/laah123graph/LAAH.

The local scanning orthogonal translation computed laminography (OTCL) has great potential for tiny fault detection of laminated structure thin-plate parts. However, it generates limited-angle and truncated projection data, which result in aliasing and truncation artifacts in the reconstructed images. The directional total variation (DTV) algorithm has been demonstrated to achieve highly accurate reconstructed images in limited-angle computed tomography (CT). However, its application in local scanning OTCL has not been explored. Based on this algorithm, we introduce the $l_p$ norm to better suppress artifacts, and prior information to further constrain the reconstructed image. Thus, we propose a prior information guided directional total p-variation (DTpV) algorithm (Pig-DTpV). The Pig-DTpV model is a constrained non-convex optimization model. The constraint term are the six DTpV terms, whereas the objective term is the data fidelity term. Then, we use the iterative reweighting strategy and the Chambolle–Pock (CP) algorithm to solve the model. The Pig-DTpV reconstruction algorithm’s performance is compared with other algorithms such as simultaneous algebraic reconstruction technique (SART), TV, reweighted anisotropic-TV (RwATV), and DTV in simulation and real data experiments. The experiment results demonstrate that the Pig-DTpV algorithm can reduce truncation and aliasing artifacts and enhance the quality of reconstructed images.

Ptychography is an imaging technique that uses the redundancy of information generated by the overlapping of adjacent light regions to calculate the relative phase of adjacent regions and reconstruct the image. In the terahertz domain, in order to make the ptychography technology better serve engineering applications, we propose a set of deep learning terahertz ptychography system that is easier to realize in engineering and plays an outstanding role. To address this issue, we propose to use a powerful deep blind degradation model which uses isotropic and anisotropic Gaussian kernels for random blurring, chooses the downsampling modes from nearest interpolation, bilinear interpolation, bicubic interpolation and down-up-sampling method, and introduces Gaussian noise, JPEG compression noise, and processed detector noise. Additionally, a random shuffle strategy is used to further expand the degradation space of the image. Using paired low/high resolution images generated by the deep blind degradation model, we trained a multi-layer residual network with residual scaling parameters and dense connection structure to achieve the neural network super-resolution of terahertz ptychography for the first time. We use two representative neural networks, SwinIR and RealESRGAN, to compare with our model. Experimental result shows that the proposed method achieved better accuracy and visual improvement than other terahertz ptychographic image super-resolution algorithms. Further quantitative calculation proved that our method has significant advantages in terahertz ptychographic image super-resolution, achieving a resolution of 33.09 dB on the peak signal-to-noise ratio (PSNR) index and 3.05 on the naturalness image quality estimator (NIQE) index. This efficient and engineered approach fills the gap in the improvement of terahertz ptychography by using neural networks.

Transformer-based light field (LF) super-resolution (SR) methods have recently achieved significant performance improvements due to global feature modeling by self-attention mechanisms. However, as a method designed for natural language processing, 4D LFs are reshaped into 1D sequences with an immense set of tokens, which results in a quadratic computational complexity cost. In this paper, a spatial–angular–epipolar swin transformer (SAEST) is proposed for spatial and angular SR (SASR), which sufficiently extracts SR information in the spatial, angular, and epipolar domains using local self-attention with shifted windows. Specifically, in SAEST, a spatial swin transformer and an angular standard transformer are firstly cascaded to extract spatial and angular SR features, separately. Then, the extracted SR feature is reshaped into the epipolar plane image pattern and fed into an epipolar swin transformer to extract the spatial–angular correlation information. Finally, several SAEST blocks are cascaded in a Unet framework to extract multi-scale SR features for SASR. Experiment results indicate that SAEST is a fast transformer-based SASR method with less running time and GPU consumption and has outstanding performance on simulated and real-world public datasets.

In order to accurately identify occluding targets and infer the motion state of objects, we propose a Bird’s-Eye View Object Detection Network based on Temporal-Spatial feature fusion (TS-BEV), which replaces the previous multi-frame sampling method by using the cyclic propagation mode of historical frame instance information. We design a new Temporal-Spatial feature fusion attention module, which fully integrates temporal information and spatial features, and improves the inference and training speed. In response to realize multi-frame feature fusion across multiple scales and views, we propose an efficient Temporal-Spatial deformable aggregation module, which performs feature sampling and weighted summation from multiple feature maps of historical frames and current frames, and makes full use of the parallel computing capabilities of GPUs and AI chips to further improve efficiency. Furthermore, in order to solve the lack of global inference in the context of temporal-spatial fusion BEV features and the inability of instance features distributed in different locations to fully interact, we further design the BEV self-attention mechanism module to perform global operation of features, enhance global inference ability and fully interact with instance features. We have carried out extensive experimental experiments on the challenging BEV object detection nuScenes dataset, quantitative results show that our method achieves excellent performance of 61.5% mAP and 68.5% NDS in camera-only 3D object detection tasks, and qualitative results show that TS-BEV can effectively solve the problem of 3D object detection in complex traffic background with lack of light at night, with good robustness and scalability.

Although there have been many previous studies on icon visual search and recognition performance, only a few have considered the effects of both the internal and external characteristics of icons. In this behavioral study, we employed a visual search task and a semantic recognition task to explore the effects of icon style, semantic distance (SD), and task difficulty on users’ performance in perceiving and identifying icons. First, we created and filtered 64 new icons, which were divided into four different groups (flat design & close SD, flat design & far SD, skeuomorphic design & close SD, skeuomorphic design & far SD) through expert evaluation. A total of 40 participants (13 men and 27 women, ages ranging from 19 to 25 years, mean age = 21.9 years, SD=1.93) were asked to perform an icon visual search task and an icon recognition task after a round of learning. Participants’ accuracy and response time were measured as a function of the following independent variables: two icon styles (flat or skeuomorphic style), two levels of SD (close or far), and two levels of task difficulty (easy or difficult). The results showed that flat icons had better visual search performance than skeuomorphic icons; this beneficial effect increased as the task difficulty increased. However, in the icon recognition task, participants’ performance in recalling skeuomorphic icons was significantly better than that in recalling flat icons. Furthermore, a strong interaction effect between icon style and task difficulty was observed for response time. As the task difficulty decreased, the difference in recognition performance between these two different icon styles increased significantly. These findings provide valuable guidance for the design of icons in human–computer interaction interfaces.