检索结果-内蒙古大学图书馆

12th International Conference on Human-computer Interaction with Mobile Devices and Services, Mobile HCI2010

作者： Ketabdar, Hamed Roshandel, Mehran Yüksel, Kamer Ali Quality and Usability Lab. Deutsche Telekom Laboratories TU Berlin Ernst-Reuter-Platz 7 10587 Berlin Germany Deutsche Telekom Laboratories Ernst-Reuter-Platz 7 10587 Berlin Germany Computer Vision and Pattern Analysis Lab. Sabanci University Istanbul Turkey

ISBN: (纸本)9781605588353

In this work, we present a new approach for text (mainly digit) entry based on digit shaped gestures created in 3D space around a mobile device. Some new mobile devices such as Apple iPhone 3GS and Google Android are equipped with magnetic (compass) sensor. The main idea is to influence the magnetic sensor using a magnet taken in hand. The user draws (writes) digits in the 3D space around the device using the magnet taken in hand. Movement of the magnet changes temporal pattern of magnetic field around the device which is sensed and registered by the magnetic (compass) sensor. The registered pattern is then compared against already recorded templates for different digits. Such a text (digit) entry approach can be especially useful for small mobile devices in which it is hard to operate small buttons or touch screen. Using our technique, the text entry space extends beyond physical boundaries of the device. A demonstrator for this approach is implemented on Apple iPhone 3GS platform. It demonstrates registering a few templates for different digits, and recognizing digits written in the space around the device.

关键词： Magnets

来源：评论

学校读者我要写书评

暂无评论

On Moving Object Segmentation from Monocular Video with Transformers

On Moving Object Segmentation from Monocular Video with Tran...

引用

International Conference on computer vision Workshops (ICCV Workshops)

作者： Christian Homeyer Christoph Schnörr Robert Bosch GmbH Corporate Research Computer Vision Lab Hildesheim Germany Image and Pattern Analysis Group Heidelberg University Germany Image and Pattern Analysis Group Heidelberg University Germany

Moving object detection and segmentation from a single moving camera is a challenging task, requiring an understanding of recognition, motion and 3D geometry. Combining both recognition and reconstruction boils down to a fusion problem, where appearance and motion features need to be combined for classification and *** this paper, we present a novel fusion architecture for monocular motion segmentation - M 3 Former, which leverages the strong performance of transformers for segmentation and multi-modal fusion. As reconstructing motion from monocular video is ill-posed, we systematically analyze different 2D and 3D motion representations for this problem and their importance for segmentation performance. Finally, we analyze the effect of training data and show that diverse datasets are required to achieve SotA performance on Kitti and Davis. Code will be released upon publication.

关键词：

来源：评论

学校读者我要写书评

暂无评论

MULTIPLE HYPOTHESES DATA ASSOCIATION PROPAGATION FOR ROBUST MONOCULAR-BASED SLAM ALGORITHMS

MULTIPLE HYPOTHESES DATA ASSOCIATION PROPAGATION FOR ROBUST ...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Mauricio Soto-Alvarez Petri Honkamaa Pattern Analysis & Computer Vision Istituto Italiano di Tecnologia (IIT) via Morego 30 16163 Genoa Italy VTT Technical Research Centre of Finland Espoo Finland

ISBN: (纸本)9781479928941

Data Association is probably the most important step of every monocular Simultaneous Localization and Mapping (SLAM) algorithm because it provides the basic information to the estimation module, independently on the estimation algorithm of choice. Although important, it is also a difficult task because the analytic solution is NP-Hard. The usual approximation is obtaining only one data association hypothesis per frame which affects the robustness of the result [1][2][3][4][5]. In this paper, a data association approach is presented, where multiple hypotheses are propagated between frames using a probabilistic framework. Experimental results, using real and synthetic data, show that the proposed algorithm produces promising results with respect to other state of the art methods.

关键词： data association state of the art supersonic low altitude missile FRAMES synthetic data simultaneous localization and mapping algorithms

来源：评论

学校读者我要写书评

暂无评论

On Moving Object Segmentation from Monocular Video with Transformers

arXiv

引用

arXiv 2024年

作者： Homeyer, Christian Schnörr, Christoph Robert Bosch GmbH Corporate Research Computer Vision Lab Hildesheim Germany Image and Pattern Analysis Group Heidelberg University Germany

Moving object detection and segmentation from a single moving camera is a challenging task, requiring an understanding of recognition, motion and 3D geometry. Combining both recognition and reconstruction boils down to a fusion problem, where appearance and motion features need to be combined for classification and segmentation. In this paper, we present a novel fusion architecture for monocular motion segmentation - M3Former, which leverages the strong performance of transformers for segmentation and multi-modal fusion. As reconstructing motion from monocular video is ill-posed, we systematically analyze different 2D and 3D motion representations for this problem and their importance for segmentation performance. Finally, we analyze the effect of training data and show that diverse datasets are required to achieve SotA performance on Kitti and Davis. Copyright © 2024, The Authors. All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Concentration estimation of ultrasound targeted microbubbles for improving cancer detection: A preliminary in VIVO study

Concentration estimation of ultrasound targeted microbubbles...

引用

IEEE International Symposium on Biomedical Imaging

作者： Claudia Sciallero Andrea Trucco Dept. of Electrical Electronic Telecommunications Engineering and Naval Architecture (DITEN) University of Genoa Genova Italy Pattern Analysis & Computer Vision Istituto Italiano di Tecnologia (IIT) Genova Italy

Ultrasound imaging can be combined with specific targeted contrast agents for noninvasive detection of pathology using disease-associated molecular signatures. The quantification of contrast agents effectively bound to receptors indicative of disease can provide additional information on the presence of pathology. This paper focuses on the opportunity of exploiting the echoes produced by targeted contrast agents to remotely estimate the volumetric concentration in a desired resolution cell. Tumor growth was induced by subcutaneous injection of prostate cancer cells in mice. A contrast-enhanced specific technique was used to image the region of interest subdivided in resolution cells of 1 mm-side. Radio frequency data were acquired for off-line processing. A method based on the support vector machine was employed to estimate the contrast agent concentration in each resolution cell. The preliminary in vivo experiments demonstrated a correlation between the extent of the primary tumor and the estimated concentrations. The combination of molecular imaging and contrast agent quantification provides the potential for earlier characterization of pathologies on molecular level.

关键词： Ultrasonic imaging Tumors Image resolution In vivo Estimation Molecular imaging

来源：评论

学校读者我要写书评

暂无评论

IFFNeRF: Initialisation Free and Fast 6DoF pose estimation from a single image and a NeRF model

IFFNeRF: Initialisation Free and Fast 6DoF pose estimation f...

引用

IEEE International Conference on Robotics and Automation (ICRA)

作者： Matteo Bortolon Theodore Tsesmelis Stuart James Fabio Poiesi Alessio Del Bue Pattern Analysis and Computer Vision (PAVIS) Fondazione Istituto Italiano di Tecnologia (IIT) Genoa Italy Technologies of Vision (TeV) Fondazione Bruno Kessler (FBK) Trento Italy University of Trento Trento Italy Department of Computer Science Durham University UK

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

We introduce IFFNeRF to estimate the six degrees-of-freedom (6DoF) camera pose of a given image, building on the Neural Radiance Fields (NeRF) formulation. IFFNeRF is specifically designed to operate in real-time and eliminates the need for an initial pose guess that is proximate to the sought solution. IFFNeRF utilizes the Metropolis-Hasting algorithm to sample surface points from within the NeRF model. From these sampled points, we cast rays and deduce the color for each ray through pixel-level view synthesis. The camera pose can then be estimated as the solution to a Least Squares problem by selecting correspondences between the query image and the resulting bundle. We facilitate this process through a learned attention mechanism, bridging the query image embedding with the embedding of parameterized rays, thereby matching rays pertinent to the image. Through synthetic and real evaluation settings, we show that our method can improve the angular and translation error accuracy by 80.1% and 67.3%, respectively, compared to iNeRF while performing at 34fps on consumer hardware and not requiring the initial pose guess. Project page: https://***/frenerf/

关键词： Accuracy Image color analysis Robot vision systems Pose estimation Memory management Neural radiance field Cameras

来源：评论

学校读者我要写书评

暂无评论

Optimizing enhanced hybrid medical imaging through dual-mode contrast agents

Optimizing enhanced hybrid medical imaging through dual-mode...

引用

IEEE International Workshop on Imaging Systems and Techniques (IST)

作者： Claudia Sciallero Luca Balbi Andrea Trucco Dept. of Electrical Electronic University of Genoa Genoa Italy Esaote S.p.A. Genoa Italy Pattern Analysis & Computer Vision Istituto Italiano di Tecnologia (IIT) Genoa Italy

Only recently, dual-mode agents able to combine ultrasound and magnetic resonance imaging have been proposed. In this paper, two types of innovative dual-mode contrast agents are investigated and compared in terms of imaging performance. Existing ultrasound contrast agents composed of gas-filled microbubbles covered by a polymeric shell were modified to obtain dual-mode contrast agents. Different densities of superparamagnetic iron oxide nanoparticles were anchored to the external bubble surface or physically embedded into the bubble shell. The goal of this paper is to individuate the setup parameters and imaging techniques able to optimize the microbubble detection with both imaging modalities. Moreover, the type of dual-mode agent and nanoparticle density that offer the best balance between magnetic resonance and ultrasound contrast enhancement is identified. In vitro imaging experiments were carried out by applying low magnetic field (0.25 T), low acoustic pressures (

关键词： Magnetic resonance imaging Ultrasonic imaging Nanoparticles Chirp Polymers Acoustics

来源：评论

学校读者我要写书评

暂无评论

LEDNet: Deep Network for Single Image Haze Removal 2018

LEDNet: Deep Network for Single Image Haze Removal

引用

Proceedings of the 11th Indian Conference on computer vision, Graphics and Image Processing

作者： Akshay Dudhane Subrahmanyam Murala Abhinav Dhall Computer Vision and Pattern Recognition Lab Indian Institute of Technology Ropar India Learning Affect and Semantic Image Analysis Group Indian Institute of Technology Ropar India

ISBN: (纸本)9781450366151

Haze during the bad weather, degrades the visibility of the scene drastically. Degradation of scene visibility varies with respect to the transmission coefficient/map (Tc) of the scene. Estimation of accurate Tc is key step to reconstruct the haze free scene. Previously, local as well as global priors were proposed to estimate the Tc. We, on the other hand, propose integration of local and global approaches to learn both point level and object level Tc. The proposed local encoder decoder network (LEDNet) estimates the scene transmission map in two stages. During first stage, network estimates the point level Tc using parallel convolutional filters and spatial invariance filtering. The second stage comprises of a two level encoder-decoder architecture which anticipates the object level Tc. We also propose, local air-light estimation (LAE) algorithm, which is able to obtain the air-light component of the outdoor scene. Combination of LEDNet and LAE improves the accuracy of haze model to recover the scene radiance. Structural similarity index, mean square error and peak signal to noise ratio are used to evaluate the performance of the proposed approach for single image haze removal. Experiments on benchmark datasets show that LEDNet outperforms the existing state-of-the-art methods for single image haze removal.

关键词： Haze removal transmission map CNN

来源：评论

学校读者我要写书评

暂无评论

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Suli Ji Baochang Zhang Dandan Du Biao He Jianzhuang Liu Science and Technology on Aircraft Control Labority Beihang University Beijing China Pattern Analysis and Computer Vision (PAVIS) Istituto Italiano di Tecnologia (IIT) Italy Chinese Department of Information Engineering Huawei Technologies Co. Ltd

ISBN: (纸本)9781479914821

This paper presents a non-linear pre-processing method based on Similarity Michaelis-Menten law (SMML) for face recognition. Similarity Michaelis-Menten law can be used to explain visual sensitivity in the vertebrate retina. We preprocess input images using SMML, and then employ Local Binary pattern (LBP) for face feature extraction. Advantages of SMML include improvement of light adaption, noise effect, detection right rate, robustness and efficiency, which inspire us exploit it for face pre-processing descriptor for the first time in the field of face recognition. And the parameters of SMML are spatiotemporally and locally estimated by the input image itself employing Sobel, which shows its advantages for face recognition. Extensive experiments clearly demonstrate the superiority of our method over the ones which only use LBP on FERET database in many aspects including the robustness against different facial expressions, lighting and aging of the subjects.

关键词： Face recognition Retina Photoreceptors Adaptation models Face Mathematical model Histograms

来源：评论

学校读者我要写书评

暂无评论

Artificial Intelligence tools to predict the level of defectiveness of existing bridges

引用

Procedia Structural Integrity 2023年 44卷 2020-2027页

作者： Agnese Natali Milind G. Padalkar Vincenzo Messina Walter Salvatore Pietro Morerio Alessio Del Bue Carlos Beltrán-González Department of Industrial and Civil Engineering University of Pisa Largo Lucio Lazzarino 1 56122 Pisa Italy Pattern Analysis and Computer Vision (PAVIS) Fondazione Istituto Italiano di Tecnologia (IIT) 16163 Genova Italy Consorzio Fabre Largo Lucio Lazzarino 1 56122 Pisa Italy

The “Italian guidelines for the maintenance of bridges” propose a qualitative method for the classification of structural, seismic, hydraulic and geotechnical risk of infrastructures. Focusing on the structural risk, one of the key parameters that significantly drives the evaluation of the vulnerability class is the level of defectiveness. The level of defectiveness can be determined only after the execution of the visual inspections, which are necessary to point out the type of damages that affect the structure, their intensity, size and position in each structural component of the bridge. Given the high number of structures to be checked and the time that is necessary to execute the visual inspections of all these bridges, an instrument to have a starting idea of the conservation status of the structure could be helpful to establish an order of priority for the bridges to be investigated. With such purpose in mind, this paper presents an ongoing activity based on the use of Artificial Intelligence to develop a smart tool that recognizes the different elements that compose the bridge, the defects, their intensity, size and position. This tool could be applied to an automatic image collection process, e.g. using a drone that, with minimal user interaction, captures images of the structure and provides quick feedback to the operators.

关键词： Bridges visual inspections Damage detection Artificial Intelligence Deep Learning

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：