检索结果-南通市图书馆

Chronological Video Synopsis via Events Rearrangement Optimization

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2018年第2期27卷 399-404页

作者： HE Yi HAN Jun SANG Nong QU Zhiguo gao changxin Huazhong University of Science and Technology Air Force Early Warning Academy

Video synopsis aims at retrieving interested events and reducing human labor on browsing long surveillance video. Traditional video synopsis methods based on energy function minimization is high computational and time consuming. Besides, the unchronological tubes shifting strategy may cause chaotic temporality and uncomfortable collisions. We propose a spatiotemporal events rearrangement optimization algorithm,which formulates the events rearrangement problem as iteration judgment on trajectory correlation and events compactability to generate the output synopsis video. An events subsection modification strategy has been presented to solve the spatial incompleteness of certain single event caused by imperfect object tracking. The experimental results have shown that the output synopsis video reserves the chronological order and avoids collisions of events.

关键词： Video synopsis Events rearrangement Events subsection modification

Semantic segmentation via pixel-to-center similarity calculation

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

CAAI Transactions on Intelligence Technology 2024年第1期9卷 87-100页

作者： Dongyue Wu Zilin Guo Aoyan Li Changqian Yu Nong Sang changxin gao National Key Laboratory of Science and Technology on Multispectral Information Processing School of Artificial Intelligence and AutomationHuazhong University of Science and TechnologyWuhanChina Meituan Inc. BeijingChina

Since the fully convolutional network has achieved great success in semantic segmentation,lots of works have been proposed to extract discriminative pixel ***,the authors observe that existing methods still suffer from two typical challenges:(i)The intra-class feature variation between different scenes may be large,leading to the difficulty in maintaining the consistency between same-class pixels from different scenes;(ii)The inter-class feature distinction in the same scene could be small,resulting in the limited performance to distinguish different classes in each *** authors first rethink se-mantic segmentation from a perspective of similarity between pixels and class *** weight vector of the segmentation head represents its corresponding semantic class in the whole dataset,which can be regarded as the embedding of the class ***,the pixel-wise classification amounts to computing similarity in the final feature space between pixels and the class *** this novel view,the authors propose a Class Center Similarity(CCS)layer to address the above-mentioned challenges by generating adaptive class centers conditioned on each scenes and supervising the similarities between class *** CCS layer utilises the Adaptive Class Center Module to generate class centers conditioned on each scene,which adapt the large intra-class variation between different *** designed Class Distance Loss(CD Loss)is introduced to control both inter-class and intra-class distances based on the predicted center-to-center and pixel-to-center ***,the CCS layer outputs the processed pixel-to-center similarity as the segmentation *** experiments demonstrate that our model performs favourably against the state-of-the-art methods.

关键词： computer vision deep neural networks image segmentation scene understanding

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

基于深度神经网络的图像风格化方法综述

模式识别与人工智能 2022年第4期35卷 333-347页

作者：涂鹏琦高常鑫桑农华中科技大学人工智能与自动化学院图像信息处理与智能控制教育部重点实验室武汉430074

图像风格化旨在通过风格化模型,将一幅图像在保持语义内容不变的同时从一种风格转换到另一种风格.鉴于深度神经网络强大的特征提取和表达能力,学者们先后提出各种基于深度神经网络的图像风格化方法.文中根据风格的定义方式,将基于深度... 详细信息

图像风格化旨在通过风格化模型,将一幅图像在保持语义内容不变的同时从一种风格转换到另一种风格.鉴于深度神经网络强大的特征提取和表达能力,学者们先后提出各种基于深度神经网络的图像风格化方法.文中根据风格的定义方式,将基于深度神经网络的图像风格化方法划分为基于参考的图像风格化方法和基于域的图像风格化方法,并对相关文献进行归纳梳理.与已有相关综述不同,文中只研究基于深度神经网络的图像风格化方法,从风格定义的角度进行详尽全面的分类.最后总结目前代表性工作在图像风格化任务常用数据集上的实验结果,分析现有方法存在的问题,并基于这些问题展望未来工作.

关键词：图像风格化深度神经网络基于域的图像风格化基于参考的图像风格化

在线全文

学校读者我要写书评

暂无评论

基于深度学习的单图像三维人体重建研究综述

华中科技大学学报（自然科学版） 2024年第5期52卷 98-122页

作者：刘乐元孙见弛高韵琪高常鑫陈靓影华中师范大学国家数字化学习工程技术研究中心湖北武汉430079 华中科技大学人工智能与自动化学院湖北武汉430074

对近年来基于深度学习的单图像三维人体重建的研究现状和发展趋势进行了总结.首先,从模型表示和计算方法两个方面梳理了当前主要的单图像三维人体重建算法.在模型表示上详细介绍了四种常见表示方式及它们之间的相互转化关系,包括深度图... 详细信息

对近年来基于深度学习的单图像三维人体重建的研究现状和发展趋势进行了总结.首先,从模型表示和计算方法两个方面梳理了当前主要的单图像三维人体重建算法.在模型表示上详细介绍了四种常见表示方式及它们之间的相互转化关系,包括深度图像与点云表示、参数化人体模型表示、体素及语义体素表示及隐式曲面函数表示.在计算方法上深入描述了基于上述四种表示方式所提出的算法,并分析了其优缺点;接着,介绍了单图像三维人体重建常用的公共数据集和客观评价指标;然后,在公共数据集上从客观指标和可视化两个角度对当前先进方法进行了评价和对比;最后,在实验结果的基础上总结了当前方法存在的问题和挑战,并展望了单图像三维人体重建未来潜在的研究方向.

关键词：三维着衣人体重建单图像三维重建深度学习点云体素参数化模型隐式曲面函数混合模型

Generic Object Recognition with Biologically-Inspired Features

在线全文

学校读者我要写书评

暂无评论

Generic Object Recognition with Biologically-Inspired Featur...

2009 Fourth International Conference on Bio-Inspired Computing:Theories and Applications(BIC-TA2009)

作者： changxin gao 1 Nong Sang 2 Jun gao 3 Lamei Zou 4 and Qiling Tang 5 Institute for Pattern Recognition and Artificial Intelligence Huazhong University of Science and Technology Wuhan 430074 China

In this paper,a set of biologically-inspired features are presented for robust object *** proposed pyramidal feature set is obtained by extracting the geometric relationship of keypoints using a set of biologically inspired templates in different *** is proposed to describe the *** paper brings together new algorithms,representations,and insights which are quite generic and may well have broader applications in computer *** proposed approach has following ***,lifetime is applied to describe the stability of the ***,the templates,which are used to extract the geometric relationships between the keypoints,are biologically inspired structure information extractors or texture information ***,the proposed approach successfully achieves an effective trade-off between generalization ability and discrimination ability for object recognition *** experimental results on object recognition demonstrate the effectiveness of the proposed method.

关键词： geometric relationship keypoints visual cortex generalization ability discrimination ability

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

深度学习实时语义分割综述

中国图象图形学报 2024年第5期29卷 1119-1145页

作者：高常鑫徐正泽吴东岳余昌黔桑农华中科技大学人工智能与自动化学院武汉430074 类脑智能系统湖北省重点实验室武汉430074 北京三快科技有限公司(美团) 北京100102

语义分割是计算机视觉领域的一项像素级别的感知任务,目的是为图像中的每个像素分配相应类别标签,具有广泛应用。许多语义分割网络结构复杂,计算量和参数量较大,在对高分辨率图像进行像素层次的理解时具有较大的延迟,这极大限制了其在... 详细信息

语义分割是计算机视觉领域的一项像素级别的感知任务,目的是为图像中的每个像素分配相应类别标签,具有广泛应用。许多语义分割网络结构复杂,计算量和参数量较大,在对高分辨率图像进行像素层次的理解时具有较大的延迟,这极大限制了其在资源受限环境下的应用,如自动驾驶、辅助医疗和移动设备等。因此,实时推理的语义分割网络得到了广泛关注。本文对深度学习中实时语义分割算法进行了全面论述和分析。1)介绍了语义分割和实时语义分割任务的基本概念、应用场景和面临问题;2)详细介绍了实时语义分割算法中常用的技术和设计,包括模型压缩技术、高效卷积神经网络(convolutional neural network,CNN)模块和高效Transformer模块;3)全面整理和归纳了现阶段的实时语义分割算法,包括单分支网络、双分支网络、多分支网络、U型网络和神经架构搜索网络5种类别的实时语义分割方法,涵盖基于CNN、基于Transformer和基于混合框架的分割网络,并分析了各类实时语义分割算法的特点和局限性;4)提供了完整的实时语义分割评价体系,包括相关数据集和评价指标、现有方法性能汇总以及领域主流方法的同设备比较,为后续研究者提供统一的比较标准;5)给出结论并分析了实时语义分割领域仍存在的挑战,对实时语义分割领域未来可能的研究方向提出了相应见解。本文提及的算法、数据集和评估指标已汇总至https://***/xzz777/Awesome-Real-time-Semantic-Segmentation,以便后续研究者使用。

关键词：实时语义分割模型轻量化高效模块设计计算机视觉深度学习

在线全文

学校读者我要写书评

暂无评论

市政污泥干化冷凝废水污染及防治措施概述

环境与发展 2019年第6期31卷 49-49,51页

作者：潘依依高长新浙江中蓝环境科技有限公司浙江温州325000 恒运市政园林建设有限公司浙江温州325000

污泥干化是污泥耦合燃煤发电前道工序,本文对污泥干化过程中产生的冷凝水水质影响因素进行分析,并提出相应污染防治措施。

关键词：市政污泥干化冷凝废水污染防治措施

Shallow sub-lakes are essential for sustaining the successful wintering of waterbirds in Poyang Lake,China

在线全文

学校读者我要写书评

暂无评论

Avian Research 2024年第2期15卷 219-228页

作者： Mengjie Lu Zhen Zhang Peng Chen changxin Xu Bin gao Luzhang Ruan School of Life Sciences Nanchang UniversityNanchang330031China Jiangxi Province Key Laboratory of Watershed Ecosystem Change and Biodiversity Center for Watershed EcologyInstitute of Life Science and School of Life SciencesNanchang UniversityNanchang330031China

For migratory waterbirds,the quality of wintering habitat is related to spring migration and successful breeding in the next *** availability of food resources in the habitat is critical and varies within water *** the water-level fluctuations in Poyang Lake have been extremely variable interannually in recent years,the wintering waterbird populations have remained relatively ***,the mechanism of maintaining the stability is worth *** study aimed to compare the distribution of vegetation and herbivorous wa-terbirds in 2015-2016 and 2016-2017,focusing on three shallow sub-lakes and one main lake *** results showed that the emergence of tubers and the growth of Carex *** a continuous food supply and habitat for wintering waterbirds with a gradual decline in the water *** sub-lakes supported almost all of the tuber-eating waterbirds(1.42-1.62×10^(5))and most geese(1.34-1.53×10^(6)).However,the main lake area,covered with Persicaria hydropiper,did not provide adequate and accessible *** resulted in almost no distribution of tuber-eating waterbirds,with only a few geese congregating in early *** results demonstrated that the shallow sub-lakes under human control provided a different environment from the main lake and are key to sustaining the successful wintering of hundreds of thousands of migratory waterbirds in Poyang ***,we recommend refining the anthropogenic management of the shallow sub-lakes to regulate the water level to ensure the carrying capacity of Poyang Lake.

关键词： Anthropogenic management Carrying capacity Seasonal lake Water level Wintering habitat

Robust Self-Triggered Min-Max Model Predictive Control for Linear Discrete-Time Systems 36

在线全文

学校读者我要写书评

暂无评论

Robust Self-Triggered Min-Max Model Predictive Control for L...