检索结果-内蒙古大学图书馆

A Programmable and Flexible Vision Processor

IEEE TRANSACTIONS ON CIRCUITS AND systems ii-ExPRESS BRIEFS 2022年第9期69卷 3884-3888页

作者： Luo, Qian Yao, Chunhe Ning, Ke Zheng, xuemin Zhao, Mingxin Cheng, Li Yu, Shuangming Liu, Jian Wu, Nanjian Liu, Liyuan Chinese Acad Sci State Key Lab Superlattices & Microstruct Inst Semicond Beijing 100083 Peoples R China Chinese Acad Sci Ctr Excellence Brain Sci & Intelligence Technol Beijing 100083 Peoples R China Univ Chinese Acad Sci Ctr Mat Sci & Optoelect Engn Beijing 100049 Peoples R China

Vision chips perform image capture and real-time intelligent image processing by integrating an imager and a vision processor on a single chip, having broad application prospects. This brief proposes a programmable and flexible vision processor with a dual-issue micro-architecture. The processor consists of a reconfigurable vector unit, a flexible memory access network, and a non-maximum suppression (NMS) block. It can efficiently implement both deep neural network (DNN) and traditional computer vision (CV) algorithms. The vector unit performs single-instruction multiple-vector (SIMV) parallel operations with reconfigurable vector width. The flexible memory access network adaptively supports multiple vector operations under different vector widths. A four-MAC processing element (PE) in the vector unit is designed to increase computational power and data reuse rate. The NMS block can speed up the object location processing of the detection networks. The chip is fabricated in a 28nm process. The experimental results show that the maximum clock frequency, peak performance, and peak energy efficiency are 600MHz, 1.2TOPS, and 2.03TOPS/W, respectively. The Mobilenet-Vl processing achieves a throughput of 404 fps under a 256x224 image size and an 87.15%(top-5) accuracy on the imageNet dataset.

关键词： Vision chip vision processor computer vision deep neural network parallel computation

来源：评论

学校读者我要写书评

暂无评论

Area reconnaissance with heterogeneous groups of autonomous systems 8

Area reconnaissance with heterogeneous groups of autonomous ...

引用

Emerging imaging and Sensing Technologies for Security and Defence ViiI 2023

作者： Tchouchenkov, I. Segor, F. Buller, A. Kollmann, M. Fraunhofer Institute of Optronics System Technologies and Image Exploitation IOSB Fraunhoferstr.1 Karlsruhe76133 Germany

ISBN: (纸本)9781510667099

In the last years Unmanned Vehicles for different environments (UxVs) have been recognized as relevant game changers and key technologies for a wide range of military and civilian applications. Even parallel deployment of heterogeneous autonomous assets as swarms or cooperating teams is no longer science fiction, but a realistic operational scenario. Their effectiveness can be significantly increased by optimization of number, capabilities and application strategy of assigned assets, particularly in time-critical tasks (e.g. the search for missing persons, disaster relief and modern warfare operations). Especially, the movements of the vehicles must be carefully managed. However, the use of small UxVs is often accompanied by limitations like short mission time and limited sensor coverage, which in turn can be compensated by the intelligent assignment of vehicles with increased autonomy in cooperating groups. An important basis for efficient and effective autonomous reconnaissance, especially in swarming or teaming scenarios, is movement optimization considering the capabilities of the heterogeneous vehicles equipped with different sensor systems operating in a combined mission. For this purpose, algorithms have been developed at Fraunhofer IOSB that enable appropriate planning and dynamic processing in very different situations for heterogeneous groups of cooperating vehicles. In the following, two exploration methods for reconnaissance missions are described in more detail and their possibilities are discussed and compared. © 2023 SPIE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Proceedings - 2023 32nd International Conference on parallel Architecture and Compilation Techniques, PACT 2023

Proceedings - 2023 32nd International Conference on Parallel...

引用

32nd International Conference on parallel Architecture and Compilation Techniques, PACT 2023

ISBN: (纸本)9798350342543

The proceedings contain 33 papers. The topics discussed include: MBAPIS: multi-level behavior analysis guided program interval selection for microarchitecture studies;automatic code generation for high-performance graph algorithms;SimplePIM: a software framework for productive and efficient processing-in-memory;Drishyam: an image is worth a data prefetcher;architecture-aware currying;PreFlush: lightweight hardware prediction mechanism for cache line flush and writeback;retargeting applications for heterogeneous systems with the tribble source-to-source framework;dynamic allocation of processor cores to graph applications on commodity servers;parallelizing maximal clique enumeration on GPUs;and HugeGPT: storing guest page tables on host huge pages to accelerate address translation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Acceleration of image processing algorithms based on a Single Board Computer and FPGA co-design 11

Acceleration of image processing algorithms based on a Singl...

引用

11th International Conference on Modern Circuits and systems Technologies (MOCAST)

作者： Kokotis, Petros Vourvoulakis, John Int Hellen Univ Dept Comp Informat & Telecommun Engn Serres Greece

ISBN: (数字)9781665467179

ISBN: (纸本)9781665467179

During recent years, various hardware platforms were developed, each one suitable for use in different kind of applications. Platforms based on FPGAs, DSPs, GPUs, Single Board Computers, microcontrollers extend processing capabilities and functionality in comparison with traditional personal computers based on a single CPU. Furthermore, co-design combines advantages from different types of processing units, rendering such architectures more attractive to researchers. In this paper, we achieve acceleration of image processing algorithms using a hardware platform based on a Raspberry Pi Single Board Computer and a custom designed FPGA HAT (Hardware Attached on Top) for RPi. The FPGA HAT consists of a Cyclone 10LP device The FPGA undertakes a computationally demanding load such as robotic vision algorithms exploiting parallelism, while the RPi can apply higher level operations such as running ROS (Robot Operating System). In order to overcome bottleneck in exchanging data between RPi and FPGA, a 16-bit parallel customized protocol was developed from scratch. The achieved transfer rate was about 50 Mbytes/sec when multi threaded software was implemented for the RPi. An image edge detector was implemented in order to verify the system performance. When only the RPi was used the processing rate was 48fps for images with resolution 512x512 pixels. RPi and FPGA co-design achieved processing rate 170fps for the same resolution images, which means an acceleration of about 350%. The proposed system was also evaluated in terms of power consumption.

关键词： fpga accelerator raspberry pi

来源：评论

学校读者我要写书评

暂无评论

parallel Block Compressive LiDAR imaging

引用

IEEE TRANSACTIONS ON COMPUTATIONAL imaging 2022年 8卷 385-396页

作者： Asmann, Andreas Mota, Joao F. C. Stewart, Brian D. Wallace, Andrew Michael STMicroelect R&D Amp Ltd Edinburgh EH3 5DA Midlothian Scotland Heriot Watt Univ Sch Engn & Phys Sci Edinburgh EH14 4AS Midlothian Scotland Heriot Watt Univ Ctr Doctoral Training Appl Photon Edinburgh EH14 4AS Midlothian Scotland

Y We propose an architecture for reconstructing depth images from raw photon count data. The architecture uses very sparse illumination patterns, making it not only computationally efficient, but due to the significant reduction in illumination density, also low power. The main idea is to apply compressive sensing (CS) techniques to block (or patch) regions in the array, which results in improved reconstruction performance, fast concurrent processing, and scalable spatial resolution. Using real and simulated arrayed LiDAR data, our experiments show that the proposed framework achieves excellent depth resolution for a wide range of operating distances and outperforms previous algorithms for depth reconstruction from photon count data in both accuracy and computational complexity. This enables eye-safe reconstruction of high-resolution depth maps at high frame rates, with reduced power and memory requirements. It is possible to sample and reconstruct a depth map in just 12 ms, enabling real-time applications at frame rates above 80 Hz.

关键词： Compressive sensing 3D image reconstruction LiDAR imaging parallelization

来源：评论

学校读者我要写书评

暂无评论

Subquadratic Kronecker Regression with applications to Tensor Decomposition 36

Subquadratic Kronecker Regression with Applications to Tenso...

引用

36th Conference on Neural Information processing systems (NeurIPS)

作者： Fahrbach, Matthew Fu, Gang Ghadiri, Mehrdad Google Res Mountain View CA 94043 USA Georgia Tech Atlanta GA USA

ISBN: (纸本)9781713871088

Kronecker regression is a highly-structured least squares problem min(x) parallel to Kx-b parallel to(2)(2), where the design matrix K = A((1)) circle times center dot center dot center dot circle times A((N)) is a Kronecker product of factor matrices. This regression problem arises in each step of the widely-used alternating least squares (ALS) algorithm for computing the Tucker decomposition of a tensor. We present the first subquadratic-time algorithm for solving Kronecker regression to a (1+epsilon)-approximation that avoids the exponential term O(epsilon(-N)) in the running time. Our techniques combine leverage score sampling and iterative methods. By extending our approach to block-design matrices where one block is a Kronecker product, we also achieve subquadratic-time algorithms for (1) Kronecker ridge regression and (2) updating the factor matrices of a Tucker decomposition in ALS, which is not a pure Kronecker regression problem, thereby improving the running time of all steps of Tucker ALS. We demonstrate the speed and accuracy of this Kronecker regression algorithm on synthetic data and real-world image tensors.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

fgSpMSpV: A Fine-grained parallel SpMSpV Framework on HPC Platforms

引用

ACM TRANSACTIONS ON parallel COMPUTING 2022年第2期9卷 1–29页

作者： Chen, Yuedan xiao, Guoqing Li, Kenli Piccialli, Francesco Zomaya, Albert Y. Hunan Univ Coll Comp Sci & Elect Engn Changsha 410082 Hunan Peoples R China Natl Supercomp Ctr Changsha Changsha 410082 Hunan Peoples R China Univ Naples Federico II Dept Elect Engn & Informat Technol I-80100 Naples Italy Univ Sydney Sch Informat Technol Sidney BC 2006 Canada

Sparse matrix-sparse vector (SpMSpV) multiplication is one of the fundamental and important operations in many high-performance scientific and engineering applications. The inherent irregularity and poor data locality lead to two main challenges to scaling SpMSpV over high-performance computing (HPC) systems: (i) a large amount of redundant data limits the utilization of bandwidth and parallel resources;(ii) the irregular access pattern limits the exploitation of computing resources. This paper proposes a fine-grained parallel SpMSpV (fgSpMSpV) framework on Sunway TaihuLight supercomputer to alleviate the challenges for large-scale real-world applications. First, fgSpMSpV adopts an MPI+OpenMP+x parallelization model to exploit the multi-stage and hybrid parallelism of heterogeneous HPC architectures and accelerate both pre-/post-processing and main SpMSpV computation. Second, fgSpMSpV utilizes an adaptive parallel execution to reduce the pre-processing, adapt to the parallelism and memory hierarchy of the Sunway system, while still tame redundant and random memory accesses in SpMSpV, including a set of techniques like the fine-grained partitioner, re-collection method, and Compressed Sparse Column Vector (CSCV) matrix format. Third, fgSpMSpV uses several optimization techniques to further utilize the computing resources. fgSpMSpV on the Sunway TaihuLight gains a noticeable performance improvement from the key optimization techniques with various sparsity of the input. Additionally, fgSpMSpV is implemented on an NVIDIA Tesal P100 GPU and applied to the breath-first-search (BFS) application. fgSpMSpV on a P100 GPU obtains the speedup of up to 134.38x over the state-of-the-art SpMSpV algorithms, and the BFS application using fgSpMSpV achieves the speedup of up to 21.68x over the state-of-the-arts.

关键词： Heterogeneous HPC manycore optimization parallelism SpMSpV

来源：评论

学校读者我要写书评

暂无评论

Neutrosophic DICOM image processing and its applications

引用

Neutrosophic Sets and systems 2023年 53卷 344-357页

作者： Nagarajan, D. Broumi, S. Department of Mathematics Rajalakshmi Institute of Technology Tamil Nadu Chennai India Laboratory of Information Processing Faculty of Science Ben M’Sik University Hassan II B.P 7955Sidi Othman Casablanca Morocco

Medical images are essential in contemporary medicine because they provide practicable entropy, which is used to diagnose medical conditions. It is useful to visualize abnormality in several parts of the body. image segmentation in the medical has an important function in various applications in diagnosis systems. Researchers have become interested in segmentation algorithms as a result of Computed Tomography (CT) and Magnetic Resonance imaging (MRI). The Region of Interest (ROI) extracts used in medical applications depend heavily on processes including cancer identification, bulk detection, and organ segmentation. Due to its capacity to deal with uncertainty and imprecision, Neutrosophic image processing (NIP) is a significant domain for uncertainty in medical image processing. Its methods in medicine demonstrate their transcendence. In the suggested work, the primary medical domains that NIP can create for image segmentation from DICOM pictures are highlighted. Due to the way it handles uncertain information, it has been found to be a better method © 2023, Neutrosophic Sets and *** Rights Reserved.

关键词： image segmentation

来源：评论

学校读者我要写书评

暂无评论

Data acquisition and data analysis in colorimetric sensor arrays

引用

CHEMOMETRICS AND INTELLIGENT LABORATORY systems 2023年 241卷

作者： Hemmateenejad, Bahram Bordbar, Mohammad Mahdi Shojaeifard, Zahra Shiraz Univ Chem Dept Shiraz Iran Shiraz Univ Med Sci Med & Nat Prod Chem Res Ctr Shiraz Iran

Responses in colorimetric sensor arrays are based on the monitoring color changes of chemical transducers by multichannel spectrophotometers or imaging recorders such as scanners, digital camera and smartphones. The images are then digitized using different image analysis algorithms. The array based methods provide a multidimensional data containing analytical, noise and outlier information. The inconsistent data can be removed or decreased by data pre-processing methods. Based on the goal of researches, the refined data can be processed by the pattern recognition methods for qualitative analysis or a by the multivariate regression models for quantitative analysis. This article explains (i) the mechanism of data collection using different types of readers, (ii) introduces the types and applications of algorithms and software for image digitalization, (iii) represents information about data preprocessing, variable selection and multivariate statistical methods and (iv) evaluates their applications in processing of colorimetric sensor array data published during the last decade.

关键词： Colorimetric sensor array Chemometrics image analysis Data acquisition Pre-processing

来源：评论

学校读者我要写书评

暂无评论

Band-Stop Smoothing Filter Design

引用

IEEE TRANSACTIONS ON SIGNAL processing 2021年 69卷 1797-1810页

作者： Roonizi, Arman Kheirati Jutten, Christian Fasa Univ Dept Comp Sci Daneshjou Blvd Fasa *** Iran Univ Grenoble Alpes Gipsa Lab CNRS F-38402 Grenoble France

Smoothness priors and quadratic variation (QV) regularization are widely used techniques in many applications ranging from signal and image processing, computer vision, pattern recognition, and many other fields of engineering and science. In this contribution, an extension of such algorithms to band-stop smoothing filters (BSSFs) is investigated. For designing a BSSF, the most important parameters are the order and the cutoff frequencies. In this paper, we show that with the optimization approaches (smoothness priors or QV regularization), the cutoff frequencies are related to the regularized parameters and the order can be directly (and easily) controlled with the number of derivatives. We describe two ways to implement the BSSFs using these approaches. First, we present a parallel structure to BSSF and then illustrate why it is less than ideal. Next, we present a novel approach regarding parallel structure to produce BSSFs with very sharp transition bands for high-performance applications. An improved optimization-based approach to BSSF design is introduced. The performance of the new BSSFs is nearly ideal.

关键词： Smoothing methods Optimization Cutoff frequency Time series analysis Signal processing algorithms Pattern recognition Linear systems Smoothness priors quadratic variation regularization least-squares optimization band-stop smoothing filter parallel structure

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：