检索结果-内蒙古大学图书馆

10th indian conference on computer vision, graphics and image processing (ICVGIP)

作者： Singh, Aditya Saini, Saurabh Shah, Rajvi Narayanan, P. J. IIIT Hyderabad KCIS CVIT Hyderabad Telangana India

ISBN: (纸本)9781450347532

User-given tags or labels are valuable resources for semantic understanding of visual media such as images and videos. Recently, a new type of labeling mechanism known as hash tags have become increasingly popular on social media sites. In this paper, we study the problem of generating relevant and useful hash-tags for short video clips. Traditional data driven approaches for tag enrichment and recommendation use direct visual similarity for label transfer and propagation. We attempt to learn a direct low-cost mapping from video to hash-tags using a two step training process. We first employ a natural language processing (NLP) technique, skip-gram models with neural network training to learn a low-dimensional vector representation of hash-tags (Tag2Vec) using a corpus of similar to 10 million hash-tags. We then train an embedding function to map video features to the low-dimensional Tag2vec space. We learn this embedding for 29 categories of short video clips with hash-tags. A query video without any tag-information can then be directly mapped to the vector space of tags using the learned embedding and relevant tags can be found by performing a simple nearest-neighbor retrieval in the Tag2Vec space. We validate the relevance of the tags suggested by our system qualitatively and quantitatively with a user study.

关键词： Tag2Vec Video Tagging Hash-tag recommendation

来源：评论

学校读者我要写书评

暂无评论

Distance Estimation for Endoscopy Based on Depth Detection in the Small Intestine Using image processing techniques 6

Distance Estimation for Endoscopy Based on Depth Detection i...

引用

6th International conference on Signal processing and Intelligent Systems (ICSPIS)

作者： Ramedani, Majid Majd, Vahid J. Tarbiat Modares Univ Dept Elect Engn Tehran Iran

ISBN: (纸本)9781728186290

In recent years, three-dimensional measurement techniques have been widely used in medical sciences, and thus, depth detection in an image plays an important role in computer vision applications. In this paper, we discuss the estimation of the distance between the head of an endoscope and the small intestine septum and its problems. the main objective is to detect the depth of the small intestine to estimate distance. images were collected through video sampling, and then the data are preprocessed. Morphological reconstruction, bounding box, Convex Hull, and Euclidean distance are employed to update the mentioned distance. At the end of this process, the outputs are simulated, and we are given the output distance in centimeters. this method will assist the endoscope to move inside the small intestine without injuries.

关键词： image processing depth detection Euclidean distance morpholog

来源：评论

学校读者我要写书评

暂无评论

Performance of Single Board computers for vision processing 11

Performance of Single Board Computers for Vision Processing

引用

IEEE 11th Annual Computing and Communication Workshop and conference (CCWC)

作者： Manore, Curtis Manjunath, Pratheek Larkin, Dominic US Mil Acad West Point NY 10996 USA

ISBN: (纸本)9781665414906

With the increasing complexity of machine vision algorithms and growing applications of image processing, how do computers without a dedicated graphics processor perform? this research discusses the computational abilities of two low-cost single board computers (SBCs) by subjecting them to various Visual Inertial Odometry (VIO) algorithms. the end goal of this research is to identify a SBC which meets the requirements of being employed on an Unmanned Aerial System for autonomous navigation.

关键词： Single Board computer Odroid computer vision VIO

来源：评论

学校读者我要写书评

暂无评论

Technology and applications of radar and tomographic medical image acquisition systems

引用

ELECTRONICS & COMMUNICATION ENGINEERING JOURNAL 1998年第3期10卷 129-138页

作者： Awcock, GJ Univ Brighton Dept Elect & Elect Engn Moulsecoomb BN2 4GJ E Sussex England

images have a very special role for humans in interpreting our world. Developments in image acquisition technology are enabling us to extend our visual faculties beyond the limitations of our physical presence and the resolving power of our eyes. this tutorial paper focuses on image acquisition technologies which give us insights that cannot be achieved at the wavelengths that we can perceive unaided. It aims to introduce two of the most rapidly developing and exciting areas in this field synthetic aperture radar imaging and tomographic medical imaging - each of which has the capacity to provide significant, but widely differing, benefits to mankind.

关键词： Data acquisition systems Nuclear medicine, emission tomography medical image processing Medical magnetic resonance imaging and spectroscopy Radiation and radioactivity applications in biomedicine biomedical NMR magnetic resonance imaging Radar equipment, systems and applications tomographic medical image acquisition systems synthetic aperture radar imaging Biomagnetism X-ray imaging positron emission tomography radar imaging Patient diagnostic methods and instrumentation data acquisition computerised tomography Biology and medical computing Optical information, image and video signal processing computer vision and image processing techniques X-rays and particle beams (medical uses) synthetic aperture radar radar

来源：评论

学校读者我要写书评

暂无评论

Unsupervised domain adaptation without source domain training samples - a maximum margin clustering based approach 16

Unsupervised domain adaptation without source domain trainin...

引用

10th indian conference on computer vision, graphics and image processing (ICVGIP)

作者： Saha, Sudipan Banerjee, Biplab Merchant, Shabbir N. Indian Inst Technol Dept Elect Engn Bombay Maharashtra India Indian Inst Technol Dept Comp Sci Roorkee Uttar Pradesh India

ISBN: (纸本)9781450347532

Unsupervised domain adaptation (DA) techniques inherently assume the presence of ample amount of source domain training samples in addition to the target domain test data. the domains are characterized by domain-specific probability distributions governing the data which are substantially different from each other. the goal is to build a task oriented classifier model that performs proportionately in both the domains. In contrary to the standard unsupervised DA setup, we propose a maximum-margin clustering (MMC) based framework for the same which does not consider source domain labeled samples. Instead we formulate it as a joint clustering problem of all the samples from both the domains in a common feature subspace. the Geodesic Flow Kernel (GFK) based subspace projection technique in the Grassmannian manifold is adopted to cast the samples in a domain invariant space. Further, the MMC stage is followed to simultaneously group the data based on the maximization of margins and a classifier is learned for each group. the data overlapping problem is taken care of by specifically learning a SVM-KNN classifier for the potentially unreliable samples per group. We validate the framework on a pair of remote sensing images of different modalities for the purpose of land-cover classification and a generic object dataset for recognition. We observe that the proposed method exhibits performances at par with the fully supervised case for both the tasks but without the requirement of costly annotations.

关键词： Domain adaptation Max-margin clustering Geodesic Flow Kernel

来源：评论

学校读者我要写书评

暂无评论

Dataset augmentation with synthetic images improves semantic segmentation 6th

Dataset augmentation with synthetic images improves semantic...

引用

6th National conference on computer vision, Pattern Recognition, image processing and graphics, NCVPRIPG 2017

作者： Goyal, Manik Rajpura, Param Bojinov, Hristo Hegde, Ravi Varanasi Varanasi221005 India Indian Institute of Technology Gandhinagar Gandhinagar382355 India Innit Inc. Redwood City94063 United States

ISBN: (纸本)9789811300196

Although Deep Convolutional Neural Networks trained with strong pixel-level annotations have significantly pushed the performance in semantic segmentation, annotation efforts required for the creation of training data remains a roadblock for further improvements. We show that augmentation of the weakly annotated training dataset with synthetic images minimizes both the annotation efforts and also the cost of capturing images with sufficient variety. Evaluation on the PASCAL 2012 validation dataset shows an increase in mean IOU from 52.80% to 55.47% by adding just 100 synthetic images per object class. Our approach is thus a promising solution to the problems of annotation and dataset collection. © Springer Nature Singapore Pte Ltd. 2018.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Performance Estimation of Numerical Biospeckle Indexing Techniques Using Rotating Diffuser Based Fixed Correlation Modelling 6

Performance Estimation of Numerical Biospeckle Indexing Tech...

引用

6th International conference on computer Applications In Electrical Engineering-Recent Advances (CERA)

作者： Chatterjee, Amit Dhanotia, Jitendra Bhatia, Vimal Prakash, Shashi Indian Inst Technol Indore Discipline Elect Engn Signals & Software Grp Indore 453552 Madhya Pradesh India Devi Ahilya Univ Inst Engn & Technol Dept Elect & Instrumentat Engn Indore 452017 Madhya Pradesh India

ISBN: (纸本)9781509048748

In this paper, we propose a simple system model to generate a temporal sequence of biospeckle patterns which is based on speckle dynamicity by clockwise rotation of a diffuser plate in coherence illumination field. these generated patterns are applied to two popular biospeckle quantification strategies namely inertia moment and absolute value of difference. Performance of both algorithms is assessed and compared by using a number of fixed correlated synthetic speckle sequences.

关键词： image processing metrology biospeckle simulation model inertia moment absolute value difference

来源：评论

学校读者我要写书评

暂无评论

Transferring Appearance with Mask-Guided Attention 24

Transferring Appearance with Mask-Guided Attention

引用

15th indian conference on computer vision graphics and image processing

作者： Bandla, Manikanta IISc Bangalore Comp Sci & Automat Bangalore Karnataka India

ISBN: (纸本)9798400710759

Despite significant advancements in large-scale text-to-image generation and text-conditioned image editing, appearance transfer remains relatively unexplored. Transferring appearance aims to transfer an object's appearance in an appearance image to an object in the structure image so that background details are preserved and accurately reflect the transferred object's characteristics. Appearance transfer has practical applications in areas like virtual try-on and e-commerce product placement. Existing methods often require fine-tuning text-to-image diffusion models or are not applicable to virtual try-on and e-commerce scenarios, which is not ideal. In this paper, we introduce a Mask-Guided attention mechanism that replaces the existing self-attention in U-net architecture of Stable diffusion [29]. this approach can be easily integrated into the Masactrl [6] framework, enabling appearance transfer without model fine-tuning and suitable for a wide range of applications. Our method uses masks of objects in images to guide the appearance transfer process, with these masks obtained from the Segment Anything Model (SAM) [17]. this integration of SAM-generated masks allows for precise object localization and more accurate appearance transfer. We have conducted comprehensive experiments on transferring various clothing items (shirts, jeans, t-shirts) onto people, as well as transferring sofas into living spaces.

关键词： Appearance transfer image editing virtual try-on sofa placement masked attention segmentation

来源：评论

学校读者我要写书评

暂无评论

A novel image encryption algorithm based on hash function

A novel image encryption algorithm based on hash function

引用

Iranian conference on Machine vision and image processing

作者： Seyedzade, Seyed Mohammad Atani, Reza Ebrahimi Mirzakuchaki, Sattar Department of Electrical Engineering Iran University of Science and Technology Tehran Iran Department of Computer Engineering University of Guilan Rasht Iran

ISBN: (纸本)9781424497065

In this paper, a novel algorithm for image encryption based on SHA-512 is proposed. the main idea of the algorithm is to use one half of image data for encryption of the other half of the image reciprocally. Distinct characteristics of the algorithm are high security, high sensitivity and high speed that can be applied for encryption of gray-level and color images. the algorithm consists of two main sections: the first does preprocessing operation to shuffle one half of image. the second uses hash function to generate a random number mask. the mask is then XORed with the other part of the image which is going to be encrypted. the aim of this work is to increase the image entropy. Both security and performance aspects of the proposed algorithm are analyzed and satisfactory results are achieved in various rounds. © 2010 IEEE.

关键词： Hash functions

来源：评论

学校读者我要写书评

暂无评论

Automatic extraction of positive cells in pathology images of meningioma based on the maximal entropy principle and HSV color space

Automatic extraction of positive cells in pathology images o...

引用

Iranian conference on Machine vision and image processing

作者： Anari, Vahid Mahzouni, Parvin Amirfattahi, Rasoul Islamic Azad University Khoramabad Branch Iran Isfehan University of Medical Science Isfahan Iran Department of Electrical and Computer Engineering Isfehan University Technology Iran

ISBN: (纸本)9781424497065

this paper describes a computer-aided system for analyzing immunohistochemically stained meningioma cancer cell images. Accurate segmentation of cells in such images plays a critical role in diagnosing diffrent type of meningioma cancer. the methodpresented to automatically extract the positive cells in meninigioma tumor immunohistochemical pathology images based on HSV color space. First, according to distribution rules of positive cells in the HSV color space, it uses the component H, S and V as threshold conditions and leverages the maximal entropy principle to build a model to segment and extract positive cells. Experimental results shows that proposed algorithm can be used by pathologist to detection reliable quantitatively analyze the parameter of tumor cells and over come to disadvantages of the traditional approach. © 2010 IEEE.

关键词： Cells

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：