检索结果-内蒙古大学图书馆

25th IEEE/RSJ International conference on Intelligent Robots and systems (IROS)

作者： Ziaeetabar, Fatemeh Kulvicius, Tomas Tamosiunaite, Minija Woergoetter, Florentin Univ Gottingen Inst Phys 3 Friedrich Hund Pl 1 D-37077 Gottingen Germany Vytautas Magnus Univ Fac Informat Kaunas Lithuania

ISBN: (纸本)9781538680940

Human-robot interaction strongly benefits from fast, predictive action recognition. For us this is relatively easy but difficult for a robot. To address this problem, here we present a novel prediction algorithm for manipulation action classes in video sequences. Manipulations are first represented using the Enriched Semantic Event Chain (ESEC) framework. This creates a temporal sequence of static and dynamic spatial relations between the objects that take part in the manipulation by which an action can be quickly recognized. We measured performance on 32 ideal as well as real manipulations and compared our method also against a state of the art trajectory-based HMM method for action recognition. We observe that manipulations can be correctly predicted after only (on average) 45% of action's total time and that we are almost twice as fast as the HMM-based method. Finally, we demonstrate the advantage of this framework in a simple robot demonstration comparing two different approaches.

关键词： Robots Three-dimensional displays Predictive models Semantics Human-robot interaction Prediction algorithms image segmentation

来源：评论

学校读者我要写书评

暂无评论

GPU based Video Object Tracking on PTZ Cameras 31

GPU based Video Object Tracking on PTZ Cameras

引用

IEEE/CVF conference on Computer Vision and Pattern Recognition (CVPR)

作者： Cigla, Cevahir Sahin, Kemal Emrecan Alim, Fikret Aselsan Inc Ankara Turkey

ISBN: (数字)9781538661000

ISBN: (纸本)9781538661000

In this study, an embedded Pan-Tilt-Zoom (PTZ) tracker system design is proposed that is based on NVIDIA Tegra K1-X1 mobile GPU platform. For this purpose, state-of-the-art correlation filter (CF) based video object tracking (VOT) algorithms are exploited regarding their high performance. Each algorithmic step is carefully implemented on GPU that further increases the efficiency and decreases execution times. The PTZ control is designed to track human targets by centralizing within the image coordinates where the targets have limited speed but obvious appearance changes. Incorporating on-board decode and encode capability of Tegra platform as well as angular position control, the presented approach enables 50-100 fps target tracking for HD (1920x1080) videos on K1 and X1 correspondingly. This is to our best knowledge the first efficient implementation of CF trackers on a mobile GPU platform with use of multiple features, scale and background adaptation. This study extends the scope of accuracy focused VOT research to platform optimized efficient implementations for real-time high resolution video tracking.

关键词： Target tracking Graphics processing units Cameras Feature extraction Correlation Real-time systems Histograms

来源：评论

学校读者我要写书评

暂无评论

Low SWaP SWIR video engine for image Intensifier replacement 44

Low SWaP SWIR video engine for Image Intensifier replacement

引用

conference on Infrared Technology and Applications XLIV

作者： Hirsh, I. Louzon, E. Aharon, A. Gazit, R. Bar, D. Kondrashov, P. Weinstein, M. Savchenko, M. Regensburger, M. Navon, A. Shunam, E. Rahat, O. Mediouni, A. Mor, E. Shay, A. Iosevich, R. Ben-Ezra, M. Tuito, A. Shtrichman, I. Semicond Devices POB 2250 IL-31021 Haifa Israel Israeli MOD Tel Aviv Israel

ISBN: (纸本)9781510617605

Night Vision Imaging in the Short-Wave Infra-Red (SWIR) has some unique advantages over Visible, Near Infra-Red (NIR) or thermal imaging. It benefits from relatively high irradiance levels and intuitive reflective imaging. InGaAs/InP is the leading technology for two-dimensional (2D) SWIR detector arrays, utilizing low dark current, high efficiency and excellent uniformity. SCD's SWIR imager is a low Size, Weight and Power (SWaP) video engine based on a low noise 640x512/15 mu m InGaAs Focal Plane Array (FPA) embedded in a low cost plastic package which includes a Thermo-Electric Cooler (TEC). The SWIR imager dimensions are 31x31x32 mm(3), it weighs 50 gram and has less than 1.4W Power consumption (excluding TEC). It supports conventional video formats, such as Camera Link and BT. 656. The video engine image processing algorithms include Non-Uniformity Correction (NUC), Auto Exposure Control (AEC), Auto Gain Control (AGC), Dynamic Range Compression (DRC) and de-noising algorithms. The algorithms are specifically optimized for Low Light Level (LLL) conditions enabling imaging from sub mlux to 100 Klux light levels. In this work we will review the optimized video engine LLL architecture, electro-optical performance and the applicability to night vision systems.

关键词： SWIR Infrared Detector InGaAs Low Light Level imaging

来源：评论

学校读者我要写书评

暂无评论

Correction of Visual Perception Based on Neuro-Fuzzy Learning for the Humanoid Robot TEO

引用

SENSORS 2018年第4期18卷 972-972页

作者： Hernandez-Vicen, Juan Martinez, Santiago Miguel Garcia-Haro, Juan Balaguer, Carlos Univ Carlos III Madrid Syst Engn & Automat Dept Avd Univ 30 Madrid 28903 Spain

New applications related to robotic manipulation or transportation tasks, with or without physical grasping, are continuously being developed. To perform these activities, the robot takes advantage of different kinds of perceptions. One of the key perceptions in robotics is vision. However, some problems related to image processing makes the application of visual information within robot control algorithms difficult. Camera-based systems have inherent errors that affect the quality and reliability of the information obtained. The need of correcting image distortion slows down image parameter computing, which decreases performance of control algorithms. In this paper, a new approach to correcting several sources of visual distortions on images in only one computing step is proposed. The goal of this system/algorithm is the computation of the tilt angle of an object transported by a robot, minimizing image inherent errors and increasing computing speed. After capturing the image, the computer system extracts the angle using a Fuzzy filter that corrects at the same time all possible distortions, obtaining the real angle in only one processing step. This filter has been developed by the means of Neuro-Fuzzy learning techniques, using datasets with information obtained from real experiments. In this way, the computing time has been decreased and the performance of the application has been improved. The resulting algorithm has been tried out experimentally in robot transportation tasks in the humanoid robot TEO (Task Environment Operator) from the University Carlos iii of Madrid.

关键词： humanoid robot artificial vision non-grasping manipulation Neuro-Fuzzy filter distortion correction

来源：评论

学校读者我要写书评

暂无评论

Real-Time Edge Template Tracking via Homography Estimation

Real-Time Edge Template Tracking via Homography Estimation

引用

25th IEEE/RSJ International conference on Intelligent Robots and systems (IROS)

作者： Qin, Xuebin He, Shida Zhang, Zichen Dehghan, Masood Jin, Jun Jagersand, Martin Univ Alberta Dept Comp Sci Edmonton AB Canada

ISBN: (纸本)9781538680940

In this paper, we propose a novel real-time method for tracking planar edge templates. This method tracks an edge template by estimating its homography transformations with respect to the sampled edge pixels detected from the incoming frames. Particularly, we define a cost function based on a new feature map of the to-be-tracked edge template and optimize it by a Lucas-Kanade-like algorithm. The feature map is defined as the fourth root of the distance transform. Our method operates on just edges so that it is good at tracking those low textured targets, such as hollow targets (mug rim), thin targets (cable, ring) and non-Lambertian objects (disc). We validate and compare our method with four other methods on five newly collected real-world video sequences. The results achieves the lowest overall average error (1.58 pixels) and also outperforms others in terms of success rate. The per frame processing time of about 30 ms proves that our method is acceptable in real-time applications. The code and dataset are publicly available at: http://***/similar to xuebin/.

关键词： image edge detection Target tracking Real-time systems Video sequences Cost function Transforms

来源：评论

学校读者我要写书评

暂无评论

Analysis of multiple features and classifier techniques combination for image pattern recognition 2nd

Analysis of multiple features and classifier techniques comb...

引用

2nd International conference on Intelligent Computing and Communication, ICICC 2017

作者： Shinde, Ashish Shinde, Abhijit Sinhgad College of Engineering PuneMaharashtra41 India Bhima Institute of Management and Technology Kagal KolhapurMaharashtra416216 India

ISBN: (纸本)9789811072444

Automatic visual pattern recognition is complex and highly researched area of image processing. This research aims to study various pattern recognition algorithms, cloth pattern recognition is presented as research problem and to find out best combination suited for the cloth pattern recognition problem. The dataset is collected from CCNY clothing pattern dataset and contains 150 samples of each category (Patternless, Striped, Plaid, and Irregular). The presented study compares all combinations of three different feature extraction techniques and three classifier techniques. Feature extraction techniques used here are Radon Feature Extraction, projection of rotated gradient, and quantized histogram of gradients. The classifiers used are KNN, neural network, and SVM classifier. The highest recognition rate is achieved using Radon Signature feature and KNN classifier combination which reaches to 93.7% of accuracy. © 2018, Springer Nature Singapore Pte Ltd.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Optimizing super-resolution reconstruction using a genetic algorithm 10

Optimizing super-resolution reconstruction using a genetic a...

引用

10th International conference on Agents and Artificial Intelligence, ICAART 2018

作者： Kawulok, Michal Kostrzewa, Daniel Benecki, Pawel Skonieczny, Lukasz Future Processing Bojkowska 37A Gliwice44-100 Poland Institute of Informatics Silesian University of Technology Akademicka 16 Gliwice44-100 Poland

ISBN: (纸本)9789897582752

Super-resolution reconstruction (SRR) is aimed at increasing spatial resolution given a single image or multiple images presenting the same scene. The existing methods are underpinned with a premise that the observed low resolution images are obtained from a hypothetic high resolution image by applying a certain imaging model (IM) which degrades the image and decreases its resolution. Hence, the reconstruction consists in applying an inverse IM to recover the high resolution data. Such an approach has been found effective, if the IM is known and controlled, in particular when the low resolution images are indeed obtained from a high resolution one. However, in a real-world scenario, when SRR is performed from images originally captured at low resolution, finding appropriate IM and tuning its hyperparameters is a challenging task. In this paper, we propose to optimize the SRR hyperparameters using a genetic algorithm, which has not been reported in the literature so far. We argue that this may substantially improve the capacities of learning the relation between low and high resolution images. Our initial, yet highly encouraging, experimental results reported in the paper allow us to outline our research pathways to deploy the developed techniques in practice. Copyright © 2018 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Black-Box Model Explained Through an Assessment of Its Interpretable Features 22nd

Black-Box Model Explained Through an Assessment of Its Inter...

引用

22nd East-European conference on Advances in Databases and Information systems (ADBIS)

作者： Ventura, Francesco Cerquitelli, Tania Giacalone, Francesco Politecn Torino Corso Duca Abruzzi 24 I-10129 Turin Italy

ISBN: (纸本)9783030000639;9783030000622

algorithms are powerful and necessary tools behind a large part of the information we use every day. However, they may introduce new sources of bias, discrimination and other unfair practices that affect people who are unaware of it. Greater algorithm transparency is indispensable to provide more credible and reliable services. Moreover, requiring developers to design transparent algorithm-driven applications allows them to keep the model accessible and human understandable, increasing the trust of end users. In this paper we present EBANO, a new engine able to produce prediction-local explanations for a black-box model exploiting interpretable feature perturbations. EBANO exploits the hypercolumns representation together with the cluster analysis to identify a set of interpretable features of images. Furthermore two indices have been proposed to measure the influence of input features on the final prediction made by a CNN model. EBANO has been preliminary tested on a set of heterogeneous images. The results highlight the effectiveness of EBANO in explaining the CNN classification through the evaluation of interpretable features influence.

关键词： Transparent mining Neural networks image processing

来源：评论

学校读者我要写书评

暂无评论

Polynomial feature engineering for classification of textural images 4

Polynomial feature engineering for classification of textura...

引用

4th International conference on Information Technology and Nanotechnology, ITNT 2018

作者： Gaidel, A.V. Samara National Research University Moskovskoe Shosse 34 Samara443086 Russia Image Processing Systems Institute Branch of the Federal Scientific Research Centre Crystallography and Photonics Russian Academy of Sciences Molodogvardeyskaya str. 151 Samara443001 Russia

I consider a number of methods of automatic quadratic features adjustment for digital textural images of biological tissues in order to improve the quality of classification. The proposed approaches are based on optimization procedures that use various quality criteria of a feature space as target functions. I investigate the methods based on random search, genetic algorithm, simulation of annealing, as well as the original hybrid algorithm. I presented results of experimental studies of the proposed algorithms on sets of real X-ray images of bone tissue and the lung CT images. We show that the hybrid algorithm provides more stable results regardless of the chosen quality criterion of the feature space, which is expressed in decreasing of the average percentage of incorrectly recognized images in comparison with the use of the specific optimization methods individually. © 2018 Institute of Physics Publishing. All rights reserved.

关键词： image enhancement

来源：评论

学校读者我要写书评

暂无评论

Simplifying the multi-GPU programming of a hyperspectral image registration algorithm

Simplifying the multi-GPU programming of a hyperspectral ima...

引用

International conference on High Performance Computing & Simulation (HPCS)

作者： Jorge Fernàndez-Fabeiro Arturo Gonzalez-Escribano Diego R. Llanos Departamento de Informaticá Universidad de Valladolid Valladolid Spain

ISBN: (数字)9781728144849

ISBN: (纸本)9781728144856

Hyperspectral image registration is a relevant task for real-time applications like environmental disasters management or search and rescue scenarios. Traditional algorithms for this problem were not really devoted to real-time performance. The HYFMGPU algorithm arose as a high-performance GPU-based solution to solve such a lack. Nevertheless, a single-GPU solution is not enough, as sensors are evolving and then generating images with finer resolutions and wider wavelength ranges. An MPI+CUDA multi-GPU implementation of HYFMGPU was previously presented. However, this solution shows the programming complexity of combining MPI with an accelerator programming model. In this paper we present a new and more abstract programming approach for this type of applications, which provides a high efficiency while simplifying the programming of the multi-device parts of the code. The solution uses Hitmap, a library to ease the programming of parallel applications based on distributed arrays. It uses a more algorithm-oriented approach than MPI, including abstractions for the automatic partition and mapping of arrays at runtime with arbitrary granularity, as well as techniques to build flexible communication patterns that transparently adapt to the data partitions. We show how these abstractions apply to this application class. We present a comparison of development effort metrics between the original MPI implementation and the one based on Hitmap, with reductions of up to 95% for the Halstead score in specific work redistribution steps. We finally present experimental results showing that these abstractions are internally implemented in a high efficient way that can reduce the overall performance time in up to 37% comparing with the original MPI implementation.

关键词： Programming Graphics processing units Hyperspectral imaging Real-time systems Libraries Principal component analysis

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：