计算机视觉论文-2021-11-03

本专栏是计算机视觉方向论文收集积累,时间:2021年9月15日,来源:paper digest

欢迎关注原创公众号 【计算机视觉联盟】,回复 【西瓜书手推笔记】 可获取我的机器学习纯手推笔记!

直达笔记地址:机器学习手推笔记(GitHub地址)

标题:知识提取的互信息估计与最大化
作者:阿曼·斯里瓦斯塔瓦;齐彦君;维森特·奥多涅斯
类别:计算机科学[计算机科学,计算机技术,数学。
亮点:在这项工作中,我们提出了互信息最大化知识蒸馏( MIM KD )。

标题:一种用于白纹伊蚊分类的深度卷积神经网络
作者:格雷齐尔阿德汗;*马赫迪德希比;戴维马西普
类别: cs.简历 [cs.简历, cs.LG |
亮点:我们介绍了应用程序的两个深度卷积神经网络在一个比较研究自动化这个分类任务。

标题:基于 YOLO 和甚深多典型相关分析的新 SAR 目标识别
作者:穆萨·阿姆拉尼;阿布德拉提夫·贝;阿布德努尔·阿马拉
类别: cs.简历 [cs.简历, cs.LG |
亮点:灵感来自非常深卷积神经网络的巨大成功(CNNs),提出了一种自适应融合不同 CNN 层有效特征的 SAR 图像目标分类鲁棒特征提取方法。

标题:PED EN et :基于块嵌入和密度估计的图像异常定位
作者:张开泰;王斌;郭丙中
类别: cs.简历 [cs.简历, cs.AI|
亮点:在这项工作中,提出了一种神经网络在无监督的图像异常定位,称为 PEDE Net 。

标题:基于片段交换的联合分段学习检索与发现
作者:沈希;艾福罗斯;朱林;马蒂厄
类别: cs.简历 [cs.简历|
亮点:这项工作的目标是有效地识别视觉上相似的模式,从一对图像,例如确定雕刻和油画之间复制的艺术品细节,或匹配的夜间照片与其白天对应。

标题:不确定度在白纹伊蚊分类中的应用
作者:格雷齐尔阿德汗;*马赫迪德希比;戴维马西普
类别: cs.简历 [cs.简历, cs.LG |
亮点:为了解决上述问题,本文提出了使用蒙特卡洛 Drop out 方法来估计的不确定性分数,以排名的分类样本,以减少需要人为监督识别白纹伊蚊。

标题:任意分辨率立体图像的神经元视差细分
作者:菲利普·阿列奥蒂等。
类别: cs.简历 [cs.简历|
亮点:我们介绍了一种新的架构,旨在促进部署的 3D 计算机视觉的廉价和广泛的消费设备,如移动电话的神经差异细化。

第八条标题:利用弹窗注释攻击视频识别模型
作者:陈凯;魏志鹏;陈菁菁;吴祖轩;蒋玉刚
类别: cs.简历 [cs.简历|
亮点:为了弥补这一差距,我们介绍了一种新的对抗性攻击在本文中,子弹屏幕注释( BSC )攻击,攻击视频识别模型与BSCs.

标题:使用像素级 AR 和PPG微弱信号的相关性
作者:毛玉茂;杨军
类别:综合类【综合类】
亮点:在本文中,我们提出了一种方案,以揭露 deep fake 通过微弱的信号隐藏在人脸视频。

第十条标题:一种基于阴影引导的生成隐式形状精确三维图像合成模型
作者:潘新钢;徐旭东;陈改来;戴博
类别: cs.简历 [cs.简历|
亮点:在这项工作中,我们提出了一种新的阴影引导的生成隐式模型,能够学习一个显着改进的形状表示,解决了这种模糊性。

第 11 条标题:基于无监督学习的单幅图像视图合成新方法
作者:刘秉政等。
类别: cs.简历 [cs.简历|
亮点:本文提出了一种无监督的网络学习这样的像素变换从一个单一的源的观点。

第 12 章,标题:基于深度亲和网络的视频监控多目标跟踪研究
作者:萨纳姆·尼萨尔·曼吉
类别: cs.简历 [cs.简历|
亮点:在本文中,最先进的MTT模型,利用深度学习的代表性权力进行审查。

第十三条标题:跨模态文本-视频检索的视觉时空关系增强网络
作者:韩宁等。
类别:cs.CV[cs.CV,cs.IR,68T07,H.3.3;H.5.1]
亮点:为了解决这个问题,我们提出了一个可视化的时空关系增强网络( VSR - Net ),一种新的跨模态检索框架,增强了组件之间的空间-时间关系的视觉表示。

第十四条标题:基于深度域竞争的无监督前景提取
作者:余培玉等。
类别: cs.简历 [cs.简历, cs.LG |
亮点:我们推出深区赛(DRC),一种以完全无监督的方式从图像中提取前景目标的算法。

第 15 条,标题:UD IS :深度视觉识别模型中偏差的无监督发现
作者:阿文库马尔·克里希那库马尔;维拉吉·帕布;斯鲁西·苏达卡尔;朱迪·霍夫曼
类别: cs.简历 [cs.简历|
亮点:在这项工作中,我们提出了 UD IS ,一个无监督的算法,用于堆焊和分析这样的故障模式。

第十六条,标题:关注视觉关键词识别
作者:普拉瓦尔;莫梅尼;阿福拉斯;齐泽曼
类别: cs.简历 [cs.简历, cs.CL|
亮点:在本文中,我们考虑的任务,在无声的视频序列中发现口语关键词-也称为视觉关键词发现。

第 17 条,标题:ST - ABN :考虑时空信息的视频识别视觉解释
作者:三原正弘;平川敏正;山下智久;藤吉平
类别: cs.简历 [cs.简历|
亮点:在本文中,我们提出了一种可视化的解释方法称为时空注意分支网络( ST - ABN )的视频识别。

第 18 条,标题:RGB 动态手势识别的多任务多模式学习
作者:范定豪;吕恒杰;徐书功;曹珊
类别: cs.简历 [cs.简历|
因此,在二维卷积神经网络的训练中,我们提出了一种端到端的多任务学习框架.

19,标题:19,标题:利用伪边缘标签的不确定性改进伪装目标检测
AUTHORS:Nobakatsu Kajiura;洪刘;Shin‘ii Satoh
类别: cs.简历 [cs.简历|
亮点:为此,我们提出了一个新的框架,充分利用多个视觉线索,即显着性和边缘,以细化预测的伪装地图。

20,标题:20,标题:弱场景下基于无线定位的无监督人再识别
刘:一成刘;文冈周;钱强熙;侯强李
类别: cs.简历 [cs.简历|
亮点:具体而言,我们提出了一种新的无监督的多模态训练框架( UM TF ),模型的视觉数据和无线信息的互补性。

第 21 条,标题:LiDAR 点云分割的假阳性检测及预测质量评价
作者:帕斯卡尔·柯林;马蒂亚斯·罗特曼;鲁茨·罗塞-科尔纳;汉诺·戈特沙尔克
分类:综合类【综合类】
亮点:我们提出了一种新的后处理工具, LiDAR 点云数据的语义分割,称为 Lidar Meta Seg ,它估计的预测质量分段。

第 22 条,标题:结合 CNN 的 Gabor 滤波器用于压缩
作者:今村明弘;日海奈奈
类别:综合类
亮点:我们遵循这个想法,并将 Gabor 滤波器的早期层CNNs为了压缩。

第二十三条,标题:钢筋信息的自动翻译GPR建筑 BIM 数据:基于深度学习的方法
作者:项中明;戈欧;阿巴斯·拉什迪
类别: cs.简历 [cs.简历|
因此,我们提出了一种连接GPR数据和 BIM 根据 FasterR-CNN.

第 24 条,标题:二维卷积神经网络在钢框架结构损伤识别中的应用
作者:沙欣·加兹维内;古拉姆雷扎·努里;赛义德·侯赛因·侯赛尼·拉瓦萨尼;瓦希德雷萨·加雷巴吉;安迪·阮
类别: cs.简历 [cs.简历, cs.LG |
亮点:在本文中,我们提出了一个应用的二维卷积神经网络( 2 - DCNNs)设计进行特征提取和分类阶段作为一个单一的有机体,以解决突出的问题。

第 25 条,标题:连续超分辨率尺度感知动态网络
作者:吴翰林;倪宁;张丽宝
类别: cs.简历 [cs.简历|
亮点:为了解决上述问题,我们提出了一个规模感知的动态网络( SA DN )的连续规模 SR 。

第 26 条标题:基于跨层比对的异构神经网络模型融合
作者:党阮;阮凯;丁奉;洪培;一浩
类别: cs.LG [cs.LG, cs.简历|
亮点:为了解决这个问题,我们提出了一个新的模型融合框架,命名为 CLA Fusion ,融合不同数量的层,我们称之为异构神经网络,通过跨层对齐的神经网络。

第 27 条,标题:潜在的认知:机器真正学会了什么
作者:皮西特·纳卡伊;吉拉代·蓬萨瓦特;塔蓬·卡坦尤库尔
类别: cs.LG [cs.LG, cs.简历|
亮点:本文探讨了一个可追溯的背景下的新解释。

第 28 条,标题:CV AD :一种基于级联 VAE 的通用医疗异常检测器
作者:郭晓元;朱慧兰;苏帕塔尔西;班纳吉
类别:第四组【第四、第五组】
亮点:我们专注于 OOD 检测的医学图像的泛化能力,并提出了一个自我监督的级联变分自编码器为基础的异常检测器( CV AD )。

第 29 条,标题:基于全容积神经网络的全脑分割
作者:李业树等。
类别:第四组【第四、第五组】
亮点:为了解决这些问题,我们建议采用一个全体积的框架,它将全体积大脑图像输入到分割网络,并直接输出整个大脑体积的分割结果。

第 30 条,标题:基于贝叶斯理论的无监督 PET 重建
作者:沈晨宇等。
类别:第四组【第四、第五组】
亮点:在本文中,我们利用DeepRED从贝叶斯的角度,在没有任何监督或辅助信息的情况下,从单个受损的正弦图中重建PET图像。

第31条,标题1:一种有效的图像恢复器:低光子计数成像的去噪与亮度调节
作者:张珊茜;林德蒙(Edmund Y.Lam)
类别:第四组【第四、第五组】
亮点:在本文中,我们研究的原始图像恢复在低光子计数条件下,通过模拟成像量子图像传感器( QIS )。

第 32 条,标题:非正式协商:无监督跨模态对抗域自适应医学图像分割框架
作者:玛丽亚巴尔德农-卡利斯托;苏珊娜赖源
类别:第四组【第四、第五组】
亮点:在这项工作中,我们提出了一个无监督的跨模态对抗域自适应( C - MA DA )框架的医学图像分割。

第 33 条,标题:19 例肺 CT 图像的人工智能语义分割与液体体积计算
作者:萨贝拉里;萨琳娜;*·伊利亚斯;尼哈·莫汉
类别:eess.IV[eess.IV,cs.CV,68T10(小学)]
亮点:这项研究的主要目的是找出GGO的体积和新冠肺炎患者的巩固,以便医生可以优先考虑病人。

34,标题:34,标题:3D-OOCS:基于感应偏置的学习前列腺分割
作者:SHRAJAN Bhandary et.艾尔。
类别:第四组【第四、第五组】
亮点:为此,我们介绍 OOCS 增强网络,一种新的架构,灵感来自于先天的性质,在脊椎动物的视觉处理。


1, TITLE: Estimating and Maximizing Mutual Information for Knowledge Distillation
AUTHORS: Aman Shrivastava ; Yanjun Qi ; Vicente Ordonez
CATEGORY: cs.CV [cs.CV, cs.IT, math.IT]
HIGHLIGHT: In this work, we propose Mutual Information Maximization Knowledge Distillation (MIMKD).

2, TITLE: A Deep Convolutional Neural Network for Classification of Aedes Albopictus Mosquitoes
AUTHORS: Gereziher Adhane ; Mohammad Mahdi Dehshibi ; David Masip
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: We introduce the application of two Deep Convolutional Neural Networks in a comparative study to automate this classification task.

3, TITLE: New SAR Target Recognition Based on YOLO and Very Deep Multi-canonical Correlation Analysis
AUTHORS: Moussa Amrani ; Abdelatif Bey ; Abdenour Amamra
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: Inspired by great success of very deep convolutional neural networks (CNNs), this paper proposes a robust feature extraction method for SAR image target classification by adaptively fusing effective features from different CNN layers.

4, TITLE: PEDENet: Image Anomaly Localization Via Patch Embedding and Density Estimation
AUTHORS: Kaitai Zhang ; Bin Wang ; C. -C. Jay Kuo
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: A neural network targeting at unsupervised image anomaly localization, called the PEDENet, is proposed in this work.

5, TITLE: Learning Co-segmentation By Segment Swapping for Retrieval and Discovery
AUTHORS: Xi Shen ; Alexei A. Efros ; Armand Joulin ; Mathieu Aubry
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: The goal of this work is to efficiently identify visually similar patterns from a pair of images, e.g. identifying an artwork detail copied between an engraving and an oil painting, or matching a night-time photograph with its daytime counterpart.

6, TITLE: On The Use of Uncertainty in Classifying Aedes Albopictus Mosquitoes
AUTHORS: Gereziher Adhane ; Mohammad Mahdi Dehshibi ; David Masip
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: In order to address the mentioned issues, this paper proposes using the Monte Carlo Dropout method to estimate the uncertainty scores in order to rank the classified samples to reduce the need for human supervision in recognising Aedes albopictus mosquitoes.

7, TITLE: Neural Disparity Refinement for Arbitrary Resolution Stereo
AUTHORS: FILIPPO ALEOTTI et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We introduce a novel architecture for neural disparity refinement aimed at facilitating deployment of 3D computer vision on cheap and widespread consumer devices, such as mobile phones.

8, TITLE: Attacking Video Recognition Models with Bullet-Screen Comments
AUTHORS: Kai Chen ; Zhipeng Wei ; Jingjing Chen ; Zuxuan Wu ; Yu-Gang Jiang
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To close this gap, we introduce a novel adversarial attack in this paper, the bullet-screen comment (BSC) attack, which attacks video recognition models with BSCs.

9, TITLE: Exposing Deepfake with Pixel-wise AR and PPG Correlation from Faint Signals
AUTHORS: Maoyu Mao ; Jun Yang
CATEGORY: cs.CV [cs.CV, eess.IV]
HIGHLIGHT: In this paper, we propose a scheme to expose Deepfake through faint signals hidden in face videos.

10, TITLE: A Shading-Guided Generative Implicit Model for Shape-Accurate 3D-Aware Image Synthesis
AUTHORS: Xingang Pan ; Xudong Xu ; Chen Change Loy ; Christian Theobalt ; Bo Dai
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this work, we address this ambiguity by proposing a novel shading-guided generative implicit model that is able to learn a starkly improved shape representation.

11, TITLE: Novel View Synthesis from A Single Image Via Unsupervised Learning
AUTHORS: BINGZHENG LIU et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This paper proposes an unsupervised network to learn such a pixel transformation from a single source viewpoint.

12, TITLE: Multi-target Tracking for Video Surveillance Using Deep Affinity Network: A Brief Review
AUTHORS: Sanam Nisar Mangi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, the state-of-the-art MTT models, which leverage from deep learning representational power are reviewed.

13, TITLE: Visual Spatio-temporal Relation-enhanced Network for Cross-modal Text-Video Retrieval
AUTHORS: NING HAN et. al.
CATEGORY: cs.CV [cs.CV, cs.IR, 68T07, H.3.3; H.5.1]
HIGHLIGHT: To solve this problem, we propose a Visual Spatio-temporal Relation-enhanced Network (VSR-Net), a novel cross-modal retrieval framework that enhances visual representation with spatio-temporal relations among components.

14, TITLE: Unsupervised Foreground Extraction Via Deep Region Competition
AUTHORS: PEIYU YU et. al.
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: We present Deep Region Competition (DRC), an algorithm designed to extract foreground objects from images in a fully unsupervised manner.

15, TITLE: UDIS: Unsupervised Discovery of Bias in Deep Visual Recognition Models
AUTHORS: Arvindkumar Krishnakumar ; Viraj Prabhu ; Sruthi Sudhakar ; Judy Hoffman
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this work, we propose UDIS, an unsupervised algorithm for surfacing and analyzing such failure modes.

16, TITLE: Visual Keyword Spotting with Attention
AUTHORS: K R Prajwal ; Liliane Momeni ; Triantafyllos Afouras ; Andrew Zisserman
CATEGORY: cs.CV [cs.CV, cs.CL]
HIGHLIGHT: In this paper, we consider the task of spotting spoken keywords in silent video sequences -- also known as visual keyword spotting.

17, TITLE: ST-ABN: Visual Explanation Taking Into Account Spatio-temporal Information for Video Recognition
AUTHORS: Masahiro Mitsuhara ; Tsubasa Hirakawa ; Takayoshi Yamashita ; Hironobu Fujiyoshi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a visual explanation method called spatio-temporal attention branch network (ST-ABN) for video recognition.

18, TITLE: Multi-Task and Multi-Modal Learning for RGB Dynamic Gesture Recognition
AUTHORS: Dinghao Fan ; Hengjie Lu ; Shugong Xu ; Shan Cao
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Therefore we propose an end-to-end multi-task learning framework in training 2D convolutional neural networks.

19, TITLE: Improving Camouflaged Object Detection with The Uncertainty of Pseudo-edge Labels
AUTHORS: Nobukatsu Kajiura ; Hong Liu ; Shin'ichi Satoh
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To this end, we propose a new framework that makes full use of multiple visual cues, i.e., saliency as well as edges, to refine the predicted camouflaged map.

20, TITLE: Unsupervised Person Re-Identification with Wireless Positioning Under Weak Scene Labeling
AUTHORS: Yiheng Liu ; Wengang Zhou ; Qiaokang Xie ; Houqiang Li
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Specifically, we propose a novel unsupervised multimodal training framework (UMTF), which models the complementarity of visual data and wireless information.

21, TITLE: False Positive Detection and Prediction Quality Estimation for LiDAR Point Cloud Segmentation
AUTHORS: Pascal Colling ; Matthias Rottmann ; Lutz Roese-Koerner ; Hanno Gottschalk
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: We present a novel post-processing tool for semantic segmentation of LiDAR point cloud data, called LidarMetaSeg, which estimates the prediction quality segmentwise.

22, TITLE: Gabor Filter Incorporated CNN for Compression
AUTHORS: Akihiro Imamura ; Nana Arizumi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We follow this idea and incorporate Gabor filters in the earlier layers of CNNs for compression.

23, TITLE: Automated Translation of Rebar Information from GPR Data Into As-Built BIM: A Deep Learning-based Approach
AUTHORS: Zhongming Xiang ; Ge Ou ; Abbas Rashidi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Thus, we propose an approach to link GPR data and BIM according to Faster R-CNN.

24, TITLE: Application of 2-D Convolutional Neural Networks for Damage Detection in Steel Frame Structures
AUTHORS: Shahin Ghazvineh ; Gholamreza Nouri ; Seyed Hossein Hosseini Lavassani ; Vahidreza Gharehbaghi ; Andy Nguyen
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: In this paper, we present an application of 2-D convolutional neural networks (2-D CNNs) designed to perform both feature extraction and classification stages as a single organism to solve the highlighted problems.

25, TITLE: Scale-Aware Dynamic Network for Continuous-Scale Super-Resolution
AUTHORS: Hanlin Wu ; Ning Ni ; Libao Zhang
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To address the above problems, we propose a scale-aware dynamic network (SADN) for continuous-scale SR.

26, TITLE: Model Fusion of Heterogeneous Neural Networks Via Cross-Layer Alignment
AUTHORS: Dang Nguyen ; Khai Nguyen ; Dinh Phung ; Hung Bui ; Nhat Ho
CATEGORY: cs.LG [cs.LG, cs.CV]
HIGHLIGHT: To address this issue, we propose a novel model fusion framework, named CLAFusion, to fuse neural networks with a different number of layers, which we refer to as heterogeneous neural networks, via cross-layer alignment.

27, TITLE: Latent Cognizance: What Machine Really Learns
AUTHORS: Pisit Nakjai ; Jiradej Ponsawat ; Tatpong Katanyukul
CATEGORY: cs.LG [cs.LG, cs.CV]
HIGHLIGHT: This article investigates the new interpretation under a traceable context.

28, TITLE: CVAD: A Generic Medical Anomaly Detector Based on Cascade VAE
AUTHORS: Xiaoyuan Guo ; Judy Wawira Gichoya ; Saptarshi Purkayastha ; Imon Banerjee
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: We focus on the generalizability of OOD detection for medical images and propose a self-supervised Cascade Variational autoencoder-based Anomaly Detector (CVAD).

29, TITLE: Whole Brain Segmentation with Full Volume Neural Network
AUTHORS: YESHU LI et. al.
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: To address these issues, we propose to adopt a full volume framework, which feeds the full volume brain image into the segmentation network and directly outputs the segmentation result for the whole brain volume.

30, TITLE: Unsupervised PET Reconstruction from A Bayesian Perspective
AUTHORS: CHENYU SHEN et. al.
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: In this article, we leverage DeepRED from a Bayesian perspective to reconstruct PET images from a single corrupted sinogram without any supervised or auxiliary information.

31, TITLE: An Effective Image Restorer: Denoising and Luminance Adjustment for Low-photon-count Imaging
AUTHORS: Shansi Zhang ; Edmund Y. Lam
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: In this paper, we investigate the raw image restoration under low-photon-count conditions by simulating the imaging of quanta image sensor (QIS).

32, TITLE: C-MADA: Unsupervised Cross-Modality Adversarial Domain Adaptation Framework for Medical Image Segmentation
AUTHORS: Maria Baldeon-Calisto ; Susana K. Lai-Yuen
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: In this work, we present an unsupervised Cross-Modality Adversarial Domain Adaptation (C-MADA) framework for medical image segmentation.

33, TITLE: AI-Powered Semantic Segmentation and Fluid Volume Calculation of Lung CT Images in Covid-19 Patients
AUTHORS: Sabeerali K. P ; Saleena T. S ; Dr. Muhamed Ilyas P ; Dr. Neha Mohan
CATEGORY: eess.IV [eess.IV, cs.CV, 68T10 (Primary)]
HIGHLIGHT: The main purpose of this study is to find the volume of GGO and consolidation of a covid-19 patient so that the physicians can prioritize the patients.

34, TITLE: 3D-OOCS: Learning Prostate Segmentation with Inductive Bias
AUTHORS: SHRAJAN BHANDARY et. al.
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: To this end, we introduce OOCS-enhanced networks, a novel architecture inspired by the innate nature of visual processing in the vertebrates.

上一篇:草稿---大


下一篇:c++ 图像尺寸不失真调整