检索结果-内蒙古大学图书馆

Convolutional neural networks for reconstruction of neutron tomography from incomplete data

NUCLEAR INSTRUMENTS & METHODS IN PHYSICS RESEARCH SECTION B-BEAM INTERACTIONS WITH MATERIALS AND ATOMS 2025年 563卷

作者： Bakirov, Bulat Kichanov, Sergey E. Kozlenko, Denis P. Joint Inst Nucl Res Frank Lab Neutron Phys Dubna 141980 Russia Russian Acad Sci Inst Archaeol Moscow 117292 Russia

To reduce the number of image projections in neutron tomography experiments, tomography reconstruction algorithms are used from an incomplete and limited number of neutron projections. The possibilities of a reconstruction algorithm based on convolutional neural networks are presented. It was found that only 72 projections are required for trained convolutional neural network for a qualitative reconstruction comparable to that from a complete dataset. Variations in the quality of the tomography reconstruction due to changes in training data and the number of input projections are considered. Examples of using convolutional neural networks for neutron tomography reconstructions of real experimental data includes the results of studies on archaeological metal materials from the "Volna-1 '' archaeological site.

关键词： CNNs Neutron imaging Neutron tomography reconstruction incomplete data Non-destructive analysis Archaeology

来源：评论

学校读者我要写书评

暂无评论

Design of an Iterative Model for Incremental Enhancements in Quantum image Processing Using Reinforcement Learning-Based Optimizations

引用

IEEE ACCESS 2025年 13卷 20491-20511页

作者： Pappala, Lalitha Kumari Veesam, Sai Babu Chattu, Kannaiah Krishna, Janapati Venkata Bodapati, Jyostna Devi Rao, B. Tarakeswara VIT AP Univ Sch Comp Sci & Engn Amaravati 522237 Andhra Pradesh India Srinivas Univ IoT & Robot Inst Engn & Technol Mangaluru 574146 India Srinivas Univ AI Dept Inst Engn & Technol Mangaluru 574146 India Srinivas Univ AR&VR Dept Inst Engn & Technol Mangaluru 574146 India Vignans Fdn Sci Technol & Res Dept Adv Comp Sci & Engn Vadlamudi 522213 Andhra Pradesh India Kallam Haranadhareddy Inst Technol Dept CSE Guntur 522019 Andhra Pradesh India

In this text we propose a collection of enhanced quantum methods that would drastically enhance the efficiency and accuracy of biomedical image processing. There are four key methodologies that have been proposed in this work: Quantum Support Vector Machines with Enhanced Feature Extraction, Quantum Generative Adversarial Networks for data Augmentation, Quantum Entanglement and Correlation Analysis for Feature Detection, and Quantum-Enhanced image reconstruction for Noisy data Samples. Q-SVM-EFE is a Quantum Support Vector Machine-based framework that incorporates advanced feature extraction to map classical image data into high-dimensional quantum states, which should capture intricate patterns and therefore be more accurate for classification. This approach scored more than 95 percent in terms of classification accuracy and tenfold increased the speed of training compared to classical SVMs. QGAN-DA generates high-quality synthetic images to alleviate the scarcity of labeled medical data. It trains a generator network through quantum circuits for the generation of new samples and a discriminator network for real image and sample classification, thus ensuring high diversity and quality in synthetic data with SSIM values above 0.90. QECA-FD applies quantum entanglement in the analysis of biomedical image correlations and is therefore very effective in detecting subtle features and anomalies. This resulted in over 95% sensitivity in anomaly detection and increased the speed of the correlation analysis by five times. QIR-ND enables enhanced image reconstruction from sullied or incomplete data through quantum noise reduction and reconstruction techniques. This new approach gave a PSNR greater than 40 dB and an eightfold improvement in reconstruction delays.

关键词： Accuracy Biomedical imaging Quantum entanglement Quantum computing Feature extraction image reconstruction Biomedical image processing Correlation Training Noise measurement generative adversarial networks image reconstruction quantum computing support vector machines

来源：评论

学校读者我要写书评

暂无评论

Analysis of Deep image Prior and Exploiting Self-Guidance for image reconstruction

IEEE TRANSACTIONS ON COMPUTATIONAL IMAGING

引用

IEEE TRANSACTIONS ON COMPUTATIONAL IMAGING 2025年 11卷 435-451页

作者： Liang, Shijun Bell, Evan Qu, Qing Wang, Rongrong Ravishankar, Saiprasad Michigan State Univ Dept Biomed Engn E Lansing MI USA Michigan State Univ Dept Math E Lansing MI 48824 USA Univ Michigan Dept Elect Engn & Comp Sci Ann Arbor MI 48901 USA Michigan State Univ Dept Computat Math Sci & Engn E Lansing MI 48824 USA Michigan State Univ Dept Biomed Engn E Lansing MI 48824 USA

The ability of deep image prior (DIP) to recover high-quality images from incomplete or corrupted measurements has made it popular in inverse problems in image restoration and medical imaging, including magnetic resonance imaging (MRI). However, conventional DIP suffers from severe overfitting and spectral bias effects. In this work, we first provide an analysis of how DIP recovers information from undersampled imaging measurements by analyzing the training dynamics of the underlying networks in the kernel regime for different architectures. This study sheds light on important underlying properties for DIP-based recovery. Current research suggests that incorporating a reference image as network input can enhance DIP's performance in image reconstruction compared to using random inputs. However, obtaining suitable reference images requires supervision and raises practical difficulties. In an attempt to overcome this obstacle, we further introduce a self-driven reconstruction process that concurrently optimizes both the network weights and the input while eliminating the need for training data. Our method incorporates a novel denoiser regularization term which enables robust and stable joint estimation of both the network input and reconstructed image. We demonstrate that our self-guided method surpasses both the original DIP and modern supervised methods in terms of MR image reconstruction performance and outperforms previous DIP-based schemes for image inpainting.

关键词： image reconstruction Electronics packaging Training Magnetic resonance imaging Noise reduction Kernel Overfitting Convolutional neural networks Inverse problems Information filters Compressed sensing deep image prior deep learning machine learning

来源：评论

学校读者我要写书评

暂无评论

Strengthening incomplete multi-view clustering: An attention contrastive learning method

引用

image AND VISION COMPUTING 2025年 157卷

作者： Hou, Shudong Guo, Lanlan Wei, Xu Anhui Univ Technol Sch Comp Sci & Technol Maanshan 243002 Anhui Peoples R China

incomplete multi-view clustering presents greater challenges than traditional multi-view clustering. In recent years, significant progress has been made in this field, multi-view clustering relies on the consistency and integrity of views to ensure the accurate transmission of data information. However, during the process of data collection and transmission, data loss is inevitable, leading to partial view loss and increasing the difficulty of joint learning on incomplete multi-view data. To address this issue, we propose a multi-view contrastive learning framework based on the attention mechanism. Previous contrastive learning mainly focused on the relationships between isolated sample pairs, which limited the robustness of the method. Our method selects positive samples from both global and local perspectives by utilizing the nearest neighbor graph to maximize the correlation between local features and latent features of each view. Additionally, we use a cross-view encoder network with self-attention structure to fuse the low dimensional representations of each view into a joint representation, and guide the learning of the joint representation through a high confidence structure. Furthermore, we introduce graph constraint learning to explore potential neighbor relationships among instances to facilitate data reconstruction. The experimental results on six multi-view datasets demonstrate that our method exhibits significant effectiveness and superiority compared to existing methods.

关键词： incomplete multi-view clustering Cross-view encoder Contrastive learning High confidence Graph constraint

来源：评论

学校读者我要写书评

暂无评论

Driving mutual advancement of 3D reconstruction and inpainting for masked faces

引用

PATTERN RECOGNITION 2025年 158卷

作者： Zhu, Guosong Qin, Zhen Zhou, Erqiang Ding, Yi Qin, Zhiguang Univ Elect Sci & Technol China Network & Data Secur Key Lab Sichuan Prov Chengdu 610054 Sichuan Peoples R China Southwestern Univ Finance & Econ Sch Comp & Artificial Intelligence Chengdu 611130 Sichuan Peoples R China

Target occlusion or pollution has always been a common and difficult problem in 3D reconstruction, seriously affecting the reconstruction effect, especially in single image scenario. To address the issues of incomplete reconstruction caused by pixel missing, a novel framework promoting mutual enhancement between 3D reconstruction and inpainting is proposed, which is capable of reconstructing a realistic 3D face with pixel completion from single image without prior assumptions. The framework is composed of an inpainting module for facial pixel completion and a reconstruction module for 3D face modeling. The inpainting module provides complete texture and accurate 3D geometry inferring premise for 3D reconstruction, and the reconstruction module is also involved in face inpainting, constraining the inpainting module to fill plausible pixels that are closer to the ground truth. Experiments show that the accurate 3D faces with complete and fine texture can be reconstructed by the proposed framework from largely masked images, with a competitive performance even surpassing most of state-of-the-art methods with unmasked images as input. Furthermore, this framework has achieved state-of-the-art performance in image inpainting on multiple face datasets.

关键词： 3D face reconstruction image inpainting

来源：评论

学校读者我要写书评

暂无评论

D3T: Dual-Domain Diffusion Transformer in Triplanar Latent Space for 3D incomplete-View CT reconstruction

引用

INTERNATIONAL JOURNAL OF COMPUTER VISION 2025年 1-24页

作者： Liu, Xuhui Li, Hong Qiao, Zhi Huang, Yawen Liu, Xi Zhang, Juan Qian, Zhen Zhen, Xiantong Zhang, Baochang Beihang Univ Beijing Peoples R China United Imaging Healthcare Cent Res Inst Beijing Peoples R China Jarvis Res Ctr Tencent YouTu Lab Beijing Peoples R China Zhongguancun Lab Beijing Peoples R China

Computed tomography (CT) is a cornerstone of clinical imaging, yet its accessibility in certain scenarios is constrained by radiation exposure concerns and operational limitations within surgical environments. CT reconstruction from incomplete views has attracted increasing research attention due to its great potential in medical applications. However, it is inherently an ill-posed problem, which, coupled with the complex, high-dimensional characteristics of 3D medical data, poses great challenges such as artifact mitigation, global incoherence, and high computational costs. To tackle those challenges, this paper introduces D3T, a new 3D conditional diffusion transformer that models 3D CT distributions in the low-dimensional 2D latent space for incomplete-view CT reconstruction. Our approach comprises two primary components: a triplanar vector quantized auto-encoder (TriVQAE) and a latent dual-domain diffusion transformer (LD3T). TriVQAE encodes high-resolution 3D CT images into compact 2D latent triplane codes which effectively factorize the intricate CT structures, further enabling compute-friendly diffusion model architecture design. Operating in the latent triplane space, LD3T significantly reduces the complexity of capturing the intricate structures in CT images. Its improved diffusion transformer architecture efficiently understands the global correlations across the three planes, ensuring high-fidelity 3D reconstructions. LD3T presents a new dual-domain conditional generation pipeline that incorporates both image and projection conditions, facilitating controllable reconstruction to produce 3D structures consistent with the given conditions. Moreover, LD3T introduces a new Dual-Space Consistency Loss that integrates image-level supervision beyond standard supervision in the latent space to enhance consistency in the 3D image space. Extensive experiments on four datasets with three inverse settings demonstrate the effectiveness of our proposal.

关键词： CT reconstruction incomplete views Triplanar auto-encoder Diffusion transformer Dual-domain conditions

来源：评论

学校读者我要写书评

暂无评论

Fast reconstruction 3D computed tomography image of stacked cell under faster scanning by dual-branch cross-fusion flat bottom network

引用

MEASUREMENT SCIENCE AND TECHNOLOGY 2025年第1期36卷 015402-015402页

作者： Chao, Long Yin, Hao Wang, Yi Tan, Chuandong Chen, Yusheng Tan, Hui Duan, Liming Chongqing Univ ICT Res Ctr Key Lab Optoelect Technol & Syst Educ Minist China Chongqing 400044 Peoples R China Chongqing Univ Coll Optoelect Engn Chongqing 400044 Peoples R China Chongqing Univ Coll Comp Sci Chongqing 400044 Peoples R China

Cone beam computed tomography (CT) fast scanning and reconstruction is a key step to achieve rapid detection of internal defects in batteries. In this work, we have achieved a faster CT scanning just in 5 s by reducing the x-ray exposure time in sparse view CT. However, the CT data is extremely incomplete by faster scanning;the existing reconstruction methods are difficult to reconstruct a high quality three-dimensional (3D) CT image of stacked cells. To address this issue, we propose a 3D CT image reconstruction network, which can reconstruct higher quality CT images from low quality 3D volume data. The input data of the reconstruction network is not 2D projection data, but 3D volume data. In this network, a high and low resolution dual-branch cross-fusion flat bottom structure is designed. The high resolution flat bottom branch aims to preserve detailed information, while the low resolution flat bottom branch focuses on capturing more semantic information. Cross-fusion between these branches mitigates the loss of semantic details. Additionally, the auxiliary loss function, the main loss function, and the 3D attention module are designed to enhance semantic accuracy and the learning performance of the network. The 3D training data is collected under a fast scanning strategy spanning 5-60 s. During the training phase, we use clipping block technology to cut the 3D volume data, enabling direct training on the 3D volume data. Our experimental results demonstrate that our 3D reconstruction network outperforms mainstream algorithms under this faster scanning strategy, which is able to reconstruct higher quality 3D CT images just in 15 s. Ablation experiments confirm the positive impact of the dual-branch cross-fusion flat bottom structure, attention module, and loss functions on improving the quality of 3D CT images.

关键词： nondestructive testing stacked cell computed tomography three-dimensional reconstruction deep learning

来源：评论

学校读者我要写书评

暂无评论

Adaptively robust high-order tensor factorization for low-rank tensor reconstruction☆

引用

PATTERN RECOGNITION 2025年 165卷

作者： Song, Zihao Chen, Yongyong Zhao, Weihua Nantong Univ Sch Math & Stat Nantong 226019 Jiangsu Peoples R China Harbin Inst Technol Shenzhen Sch Comp Sci & Technol Shenzhen 518055 Peoples R China

Recently, various approaches have been proposed for tensor reconstruction from incomplete and contaminated data. However, most algorithms focus on third-order tensors, neglecting higher-order tensors that are common in real-world applications. Additionally, many studies use LASSO-type penalties or second-order statistics to capture noise patterns, which may not perform well with dense and gross outliers. To address these challenges, we propose a novel robust high-order tensor recovery model that simultaneously removes complex noise and completes missing entries. We introduce a factor Frobenius norm for the low-rank structures of high-order tensors and derive a nonconvex function via the L2 criterion. An estimation algorithm is developed using the alternating minimization method. Our method jointly estimates tensor terms of interest and precision parameters, adapting to noise patterns for data-driven robustness. We analyze the convergence properties of our algorithm, and numerical experiments validate its superiority in natural image reconstruction, video restoration, and background modeling compared to state-of-the-art methods.

关键词： data-driven robust Robust high-order tensor reconstruction High-order tensor factor frobenius norm Low-rank Nonconvex optimization L2 criterion

来源：评论

学校读者我要写书评

暂无评论

MAGS: Max-Gap Loss-Guided Siamese-reconstruction Network for Hyperspectral image Partial Label Learning

引用

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 2025年 63卷

作者： Tian, Xiaoyu Luo, Fulin Fu, Chuan Guo, Tan Du, Bo Chanussot, Jocelyn Chongqing Univ Coll Comp Sci Chongqing 400044 Peoples R China Chongqing Univ Posts & Telecommun Sch Commun & Informat Engn Chongqing 400065 Peoples R China Wuhan Univ Sch Comp Sci Wuhan 430072 Peoples R China Univ Grenoble Alpes Inria CNRS Grenoble INPLJK F-38000 Grenoble France

Due to the powerful feature extraction capabilities of deep learning, a series of deep learning-based methods for hyperspectral image (HSI) classification have been proposed and achieved satisfactory performance. However, most of these methods require a large number of labeled data, and the collection of completely accurate pixel-level labeled HSI data is difficult, resulting from the intricate label ambiguity of HSI and incomplete prior knowledge of annotators. Simultaneously, a few researchers focus on label ambiguity for HSI classification. Partial label learning (PLL) is one of the strategies to solve the problem where each training instance is assigned a candidate label set, among which only one is the ground truth label, which can essentially alleviate labeling difficulties. In this article, a max-gap loss-guided Siamese reconstruction network (MAGS) is proposed to combine PLL with HSI classification. MAGS consists of three components, including a spatial-spectral encoder, a spatial-spectral decoder, and a Siamese spatial-spectral encoder for high-quality feature representation learning to facilitate label disambiguation. In the encoding process, MAGS introduces the cross-attention and max-matching fusion strategies to obtain more representative features. In addition, to improve label disambiguation, the maximum gap loss is designed to guide the model training. Quantitative and qualitative results indicate that the MAGS outperforms several state-of-the-art methods on three HSI datasets. The code is available at https://***/Nemo96yu/MAGS.

关键词： Feature extraction Phase locked loops Training Hyperspectral imaging data mining Representation learning Labeling Electronic mail Recurrent neural networks image reconstruction Hyperspectral image (HSI) classification label disambiguation partial label learning (PLL) spatial-spectral information

来源：评论

学校读者我要写书评

暂无评论

A Spatial Transformation Based Next Frame Predictor

引用

IEEE ACCESS 2025年 13卷 5406-5423页

作者： Mokssit, Saad Licea, Daniel Bonilla Guermah, Bassma Ghogho, Mounir Int Univ Rabat Coll Engn & Architecture TICLab Rabat 11103 Morocco Mohammed VI Polytech Univ Coll Comp Ben Guerir 43150 Morocco Univ Leeds Fac Engn Leeds LS2 9JT England

In recent years, the automobile industry has achieved astonishing success in making autonomous cars safer, more affordable, and more reliable. However, current autonomous driving technology is mainly based on reactive controllers that attempt to respond to the various events the car encounters. Yet, achieving a truly safe and reliable autonomous system necessitates anticipating such events and planning the correct actions in advance to avoid undesirable behavior. Recent advances in deep learning have shown remarkable performance in predicting future frames from video sequences. However, most of these approaches can only handle a few moving elements in the scene and perform poorly when the camera is in motion. This is mainly due to the difficulty of disentangling camera intrinsic motion from object-dependent motion. In this work, we equip autonomous cars with an object-oriented next-frame predictor that leverages Transformer architecture to extract, for each moving object in the scene, a spatial transformation applied to the object to predict its configuration in the next frame. Static elements of the scene are then used to estimate camera intrinsic motion, which is applied to the background to predict how it will be viewed in the next frame. Notably, our approach significantly reduces the complexity typically associated with such models by requiring the estimation of only 14 parameters per moving object, independent of image resolution. We have validated the generalization capabilities of our model through training on simulated datasets and testing on real-world datasets. The results indicate that our model not only outperforms existing models trained solely on real data but also exhibits superior resilience to occlusions and incomplete data in the input sequences. These findings underscore the potential of our model to significantly improve the predictive analytics capabilities of autonomous driving systems, thereby enhancing their safety and reliability in dynamic

关键词： Next frame prediction moving camera image reconstruction deep learning projective and affine transforms Next frame prediction moving camera image reconstruction deep learning projective and affine transforms

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：