检索结果-内蒙古大学图书馆

5th International conference on Computer Vision, image and Deep Learning, CVIDL 2024

作者： Yang, Yuxian Zhang, Jun Guangdong University of Technology School of Information Engineering Guangzhou China

ISBN: (纸本)9798350373820

Deep learning methods have exhibited remarkable performance in addressing various inverse problems. Nevertheless, most of existing models are trained for specific sampling processes. Their performance deteriorates significantly when the sampling process of the problem stands even slight changes. Moreover, these networks often overlook the long-distance dependencies inherent in image. In this paper, we proposed an innovative image restoration approach adeptly exploits and incorporates Non-local Prior information while accommodating Arbitrary Sampling processes, called NPASNet. Specifically, utilizing the operator decoupling property of the ADMM framework, the NPASNet separates the solving process into two stages: initially estimating rough restored signals from sampled data and subsequently projecting these signals onto more precise unknown signals based on prior information. To this end, we employ a generative adversarial network (GAN) augmented with self-attention mechanisms to extract non-local priors from images. Interestingly, the proposed NPASNet can be trained once to address a broad spectrum of image restoration tasks. Our experiments indicate that the NPASNet surpasses the existing models by a notable average of 0.5 dB, demonstrating comparable or even superior performance relative to state-of-the-art methods when confronted with compressive sensing problems. © 2024 IEEE.

关键词： image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Pred-NBV: Prediction-guided Next-Best-View Planning for 3D Object reconstruction

Pred-NBV: Prediction-guided Next-Best-View Planning for 3D O...

引用

IEEE/RSJ International conference on Intelligent Robots and Systems (IROS)

作者： Dhami, Harnaik Sharma, Vishnu D. Tokekar, Pratap Univ Maryland Dept Comp Sci College Pk MD 20742 USA

ISBN: (纸本)9781665491907

Prediction-based active perception has shown the potential to improve the navigation efficiency and safety of the robot by anticipating the uncertainty in the unknown environment. The existing works for 3D shape prediction make an implicit assumption about the partial observations and therefore cannot be used for real-world planning and do not consider the control effort for next-best-view planning. We present Pred-NBV, a realistic object shape reconstruction method consisting of PoinTr-C, an enhanced 3D prediction model trained on the ShapeNet dataset, and an information and control effort-based next-best-view method to address these issues. Pred-NBV shows an improvement of 25.46% in object coverage over the traditional methods in the AirSim simulator, and performs better shape completion than PoinTr, the state-of-the-art shape completion model, even on real data obtained from a Velodyne 3D LiDAR mounted on DJI M600 Pro.

关键词： image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Deep Learning techniques for reconstruction on ASTRI Mini-Array Monte Carlo data 38

Deep Learning techniques for reconstruction on ASTRI Mini-Ar...

引用

38th International Cosmic Ray conference, ICRC 2023

作者： Lombardi, S. Visconti, F. Mastropietro, M. INAF Osservatorio Astronomico di Roma Via Frascati 33 Monte Porzio Catone RomaI-00078 Italy ASI Space Science Data Center Via del Politecnico s.n.c. RomaI-00133 Italy

The interaction of gamma rays and cosmic rays with the Earth’s atmosphere initiate air showers that, in turn, induce the emission of Cherenkov photons detectable by ground-based Imaging Atmospheric Cherenkov Telescopes (IACTs). Any data analysis software for gamma-ray astronomy with IACTs requires an essential component to discriminate the nature of the primary particle, as well as to reconstruct its energy and arrival direction. In this field, the standard reconstruction approach is to use supervised machine learning techniques, mostly based on decision trees or Random Forest, which build models by training on simulated data using image and stereoscopic parameters as input features. This approach can be overcome by deep learning techniques, directly operating on pixelated camera images recorded by the array telescopes as input to models. In this way, all available information per each shower image can potentially be exploited for reconstruction, without relying solely on derived parameters. We evaluated some deep learning techniques on Monte Carlo simulated data of the ASTRI Mini-Array, an array of nine dual-mirror 4-m class IACTs under deployment at the Observatorio del Teide (Tenerife, Spain), sensitive to gamma-ray radiation in the 1–200 TeV energy range. In this contribution we present how deep learning algorithms such as convolutional neural networks can be used to reconstruct events acquired by the ASTRI Mini-Array;we will first describe the analysis work flow and introduce the architectures, and then compare the performance obtained with the new reconstruction methods with that of standard method. © Copyright owned by the author(s) under the terms of the Creative Commons.

关键词： Gamma rays

来源：评论

学校读者我要写书评

暂无评论

Time Dependent image Generation of Plants from incomplete Sequences with CNN-Transformer 44th

Time Dependent Image Generation of Plants from Incomplete Se...

引用

44th DAGM German conference on Pattern Recognition (DAGM GCPR)

作者： Drees, Lukas Weber, Immanuel Russwurm, Marc Roscher, Ribana Tech Univ Munich Data Sci Earth Observat Munich Germany Univ Bonn Remote Sensing IGG Bonn Germany AB EX PLEdoc Essen Germany Ecole Polytech Fed Lausanne EPFL ECEO Lab Lausanne Switzerland

ISBN: (纸本)9783031167881;9783031167874

data imputation of incomplete image sequences is an essential prerequisite for analyzing and monitoring all development stages of plants in precision agriculture. For this purpose, we propose a conditional Wasserstein generative adversarial network TransGrow that combines convolutions for spatial modeling and a transformer for temporal modeling, enabling time-dependent image generation of above-ground plant phenotypes. Thereby, we achieve the following advantages over comparable data imputation approaches: (1) The model is conditioned by an incomplete image sequence of arbitrary length, the input time points, and the requested output time point, allowing multiple growth stages to be generated in a targeted manner;(2) By considering a stochastic component and generating a distribution for each point in time, the uncertainty in plant growth is considered and can be visualized;(3) Besides interpolation, also test-extrapolation can be performed to generate future plant growth stages. Experiments based on two datasets of different complexity levels are presented: Laboratory single plant sequences with Arabidopsis thaliana and agricultural drone image sequences showing crop mixtures. When comparing TransGrow to interpolation in image space, variational, and adversarial autoencoder, it demonstrates significant improvements in image quality, measured by multi-scale structural similarity, peak signal-to-noise ratio, and Frechet inception distance. To our knowledge, TransGrow is the first approach for time- and image-dependent, high-quality generation of plant images based on incomplete sequences.

关键词： data imputation Transformer Positional encoding image time series Conditional GAN image generation Plant growth modeling

来源：评论

学校读者我要写书评

暂无评论

Preparation of a Training dataset for Reconstructing Three-dimensional Structures Using a Single Two-dimensional image 5

Preparation of a Training Dataset for Reconstructing Three-d...

引用

5th IEEE International conference on Civil Aviation Safety and Information Technology, ICCASIT 2023

作者： Xu, Geling Gao, Mingliang Luo, Ai Wang, Diao Xu, Zongshun Northwest Minzu University Gansu Engineering Research Center for Eco-Environmental Intelligent Networking College of Electrical Engineering Lanzhou China

ISBN: (纸本)9798350310603

The construction of three-dimensional porous media structures using generative adversarial network models is currently a hot research topic. When using this method for 3D reconstruction, a rich training data is required, and the quality of the database seriously affects the quality of the final 3D reconstruction results. In practical situations, sometimes only a limited number of two-dimensional slice images can be obtained. Therefore, how to use the limited two-dimensional slice images to construct the training data required for generating adversarial network models has important practical significance and research value. Based on this, this paper focuses on how to prepare a training data for reconstructing three-dimensional structures in the presence of only a single two-dimensional image. There are two stages used in our method. In the first stage, digital techniques such as cropping, bilinear interpolation, corrosion, and dilation are applied to the obtained images to enrich our database. In the second stage, the optimal proportion of images obtained from each part was found through extensive experiments, resulting in the best performance of reconstructing the three-dimensional structure. In the final part of this article, a set of core samples were used to validate the above method. The experiment showed that the proposed method of constructing a 3D reconstruction training data using a single 2D image was effective. © 2023 IEEE.

关键词： Corrosion

来源：评论

学校读者我要写书评

暂无评论

Compressive Sensing Based Algorithms for Limited-View PAT image reconstruction

Compressive Sensing Based Algorithms for Limited-View PAT Im...

引用

Asia-Pacific-Signal-and-Information-Processing-Association Annual Summit and conference (APSIPA ASC)

作者： John, Mary Josy Barhumi, Imad United Arab Emirates Univ Coll Engn Al Ain U Arab Emirates

ISBN: (纸本)9798350300673

Limited-view sensor arrangement is a major concern in medical imaging as it limits the data that the sensor could acquire. However, this limitation, signal sparsity, can be exploited using compressive sensing (CS) techniques to reconstruct high-resolution images. The objective of this research paper is to develop CS-based algorithms for reconstructing images in limited-view photoacoustic tomography. Various CS reconstruction algorithms and sensor arrangements were assessed to identify the optimal approach for reconstructing images from limited-view sensor data. The results show that the split Bregman total variation (SBTV)-l(1) CS algorithm is the most efficient for all sensor arrangements. The study also reveals that the convex sensor array yields the best results among all sensor arrangements. Additionally, the implementation of SBTV-l(1) using Cholesky factorization requires less computation time and is 10 to 15 times faster than the direct implementation.

关键词： Photoacoustic Tomography Compressive Sensing Split Bregman Limited view

来源：评论

学校读者我要写书评

暂无评论

SA-CCSNet: Saliency-Aware Cascade Network for image Compressed Sensing 6

SA-CCSNet: Saliency-Aware Cascade Network for Image Compress...

引用

6th IEEE International conference on Pattern Recognition and Artificial Intelligence, PRAI 2023

作者： Xu, Dan Li, Wanrong Shi, Jinlong Feng, Bing Jiangsu University of Science and Technology Department of Computer Science Zhenjiang China

ISBN: (纸本)9798350325485

Sampling matrix design and reconstruction scheme development are two critical issues in image compressed sensing (CS). For the first issue, uniform sampling matrices are commonly used, which ignore the characteristics of the image and lead to distortion in important image parts. For the second problem, traditional iterative optimization algorithms often suffer from high computational complexity, while deep learning-based non-iterative methods need more insight from the CS domain. To solve the two problems, we propose a saliency-Aware cascade framework that includes three serial subnetworks, i.e., sampling subnetwork, initial recovery subnetwork, and deep recovery subnetwork. There are two branches, full image sampling and salient image sampling, in the sampling subnetwork to allocate more sampling resources for salient regions. Initial recovery is used to obtain a preliminary recovered image, and deep recovery subnetwork maps the traditional Iterative Shrinkage-Thresholding Algorithm (ISTA) to CNN convolutional modules for further optimization. The entire network is learned from data end-To-end using a hybrid loss. Experimental results show that our approach results in superior quality reconstruction images and higher metrics compared to alternative algorithms, particularly at extremely low sampling rates ranging from 0.01 to 0.25. © 2023 IEEE.

关键词： Compressed sensing

来源：评论

学校读者我要写书评

暂无评论

Hybrid event-enhanced image de-occlusion 6

Hybrid event-enhanced image de-occlusion

引用

6th conference on Frontiers in Optical Imaging and Technology: Applications of Imaging Technologies

作者： Gao, Ning Huang, Feice Zhang, Lei Luo, Xiaoyan Deng, Yue School of Astronautics Beihang University Beijing100191 China

ISBN: (数字)9781510679757

ISBN: (纸本)9781510679740

Removing dense foreground occlusion from images and reconstructing the target of interest is a critical vision task. In previous studies, it was generally tackled through frame-based methods, but the performance was limited due to the lack of valid information. With the development of event cameras, their advantages in high temporal resolution and asynchronous response mechanism at each pixel have shown significant potential in various visual tasks. However, the event stream is plagued by multiple noise factors and static perceptual limitations, making it difficult to directly restore the local texture and absolute color of occluded objects. To overcome these challenges, we incorporate event stream information into the image frame restoration process to achieve a more effective occlusion removal. Specifically, we introduce a hybrid neural network for removing foreground occlusions from event-frame inputs, along with the design of an event stream encoder based on Spiking Neural Networks (SNN) and a Temporal Channel Attention Block (TCA) to enhance frame features. In addition, in order to significantly enhance the capability of occlusion removal, we introduce a General Restoration Block (GRB), which is applicable to both event data and frame data. Extensive experimental results indicate that the proposed method performs favorably against the state-of-the-art approaches. © 2024 SPIE.

关键词： image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Application of Multi-Source Remote Sensing data Fusion in Rural Sewage Treatment Effect Evaluation 2

Application of Multi-Source Remote Sensing Data Fusion in Ru...

引用

2nd IEEE International conference on Sensors, Electronics and Computer Engineering, ICSECE 2024

作者： Xu, Lansheng Zhang, Jun West Yunnan University Yunnan Lincang China

ISBN: (纸本)9798350373646

A remote-control system based on the integration of automation technology and Internet of Things technology is designed to realize the remote control, monitoring and maintenance of sewage treatment equipment. This project combines information from multiple sources under the condition of invariance. Then the universal remote sensing objects are networked detected. At the image level, the multi- source image mode transfer network uses the style transfer network to generate quasi-multiple images similar to the real space in order to shorten the distance between multiple source images, and from the image level invariant learning. In view of the invariability of multi-source data at the feature level, an adaptive network is used to decouple the information among multiple attributes. The attribute reconstruction is completed by adjusting the weight of domain attention network. It can better solve the problems of low operation and management efficiency and high operation and maintenance cost in China's rural domestic sewage treatment work. © 2024 IEEE.

关键词： Sewage treatment

来源：评论

学校读者我要写书评

暂无评论

Multi-View Supervision for Single-View reconstruction via Differentiable Ray Consistency

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2022年第12期44卷 8754-8765页

作者： Tulsiani, Shubham Zhou, Tinghui Efros, Alexei A. Malik, Jitendra Univ Calif Berkeley Dept Elect Engn & Comp Sci Berkeley CA 94720 USA

We study the notion of consistency between a 3D shape and a 2D observation and propose a differentiable formulation which allows computing gradients of the 3D shape given an observation from an arbitrary view. We do so by reformulating view consistency using a differentiable ray consistency (DRC) term. We show that this formulation can be incorporated in a learning framework to leverage different types of multi-view observations e.g., foreground masks, depth, color images, semantics etc. as supervision for learning single-view 3D prediction. We present empirical analysis of our technique in a controlled setting. We also show that this approach allows us to improve over existing techniques for single-view reconstruction of objects from the PASCAL VOC dataset.

关键词： Three-dimensional displays Shape image reconstruction Cameras Solid modeling Color Training data 3D reconstruction multi-view supervision ray consistency

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：