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:天文学中深度学习算法的鲁棒性——星系形态学研究
作者:阿依普里亚诺维奇等。
类别: 天体 ph.GA [天文 ph.加, cs.简历, cs.LG |
亮点:深度学习模型正越来越多地被广泛应用于科学领域,特别是处理高维和大量的科学数据。
标题:基于嵌入互相关的全卷积网络的单像素粒子图像测速
作者:齐高等。
分类:物理学.流感病毒【物理学.流感病毒,艾滋病病毒, CV 病毒】
亮点:在目前的工作中,我们提出了一种新的速度场估计范式,实现了深度学习方法和传统的互相关方法的协同结合。
标题:Deep Dose Net :用于放射治疗中三维剂量预测的深度学习模型
作者:穆塔兹·侯赛因·索姆罗;维克托·加布里埃尔·莱安德罗·阿尔维斯;哈米德雷扎·努尔扎德;杰弗里· V ·西伯斯
分类:物理学,医学,医学cs.LG,第四章]
亮点:提出了基于 ResNet 和 Dilated Dense Net 的 Deep Dose Net 三维剂量预测模型。
标题:从面部到步态:从行走模式中弱监督学习性别信息
作者:安迪·卡特鲁纳;阿德里安·科斯马;扬·艾米连·拉迪
类别: cs.简历 [cs.简历|
亮点:我们提出了一个弱监督的方法来学习性别信息的人的行走方式的基础上。
标题:用于三维人体姿态估计的高阶隐式光顺网络
作者:全建宁;本·哈姆扎
类别: cs.简历 [cs.简历|
亮点:在本文中,我们介绍了一种高阶图卷积框架与初始残留连接的二维到三维姿态估计。
标题:基于类级梯度操作的广义数据加权
作者:陈灿等。
类别:分类: CVcs.LG,数学网]
亮点:为此,在本文中,我们提出了广义数据加权(GDW)通过在类级别上操作梯度,同时减轻标签噪声和类不平衡。
标题:基于折线的生成式导航空间分割算法
作者:陈正;丁正明;克兰德尔;刘兰涛
类别:综合类【综合类】
亮点:在这项工作中,我们将视觉导航空间分割作为一个场景分解问题,并提出了多段线分割变分自动编码器网络( PSV - Net ),一个representation-learning-based框架,使机器人学习的导航空间分割在一个无监督的方式。
第八条标题:VPF Net :用于多类三维目标检测的体素-像素融合网络
作者:王嘉宏;陈学伟;傅立臣
分类:综合类【综合类】cs.LG,中华民国】
亮点:为了提高在复杂环境中的检测性能,本文提出了一种基于深度学习( DL )嵌入式融合的多类三维物体检测网络,既可以容纳 LiDAR 和摄像机传感器数据流,称为 Voxel - Pixel 融合网络( VPF Net )。
标题:三维表面感知图像合成中的生成占用域
作者:徐旭东;潘新刚;林大花;戴波
类别: cs.简历 [cs.简历, cs.AI|
亮点:在本文中,我们提出了生成占用领域( GOF ),一种新的模型的基础上生成的辐射场,可以学习紧凑的对象表面,而不妨碍其训练收敛。
第十条标题:DFC A Net :面向跨域虹膜呈现攻击检测的稠密特征标定-注意引导网络
作者:高拉夫·贾斯瓦尔;阿曼·维尔马;苏曼特拉·杜塔·罗伊;拉加文德拉·拉玛钱德拉
类别: cs.简历 [cs.简历|
亮点:为了缓解这些缺点,本文提出了 DFC A Net :密集的特征校准和注意力引导网络,校准本地传播的虹膜模式与全球定位的。
第 11 条标题:PP - Pico Det :一种更好的移动设备实时目标检测器
作者:余光华等。
类别: cs.简历 [cs.简历|
亮点:在这项工作中,我们致力于研究目标检测的关键优化和神经网络架构的选择,以提高准确性和效率。
第 12 章,标题:DP NET :轻量级自注意高效目标检测的双路径网络
作者:施慧敏;周泉;倪英豪;吴晓富;拉泰基
类别: cs.简历 [cs.简历|
亮点:为了解决之间的权衡计算成本和检测精度,本文提出了一种双路径网络,命名为 DP Net ,高效的对象检测轻量级的自我关注。
第十三条标题:结构信息是关键:三维目标检测中的自关注 RoI 特征提取器
作者:张殿坤;郑志杰;毕雪婷;刘晓军
类别: cs.简历 [cs.简历|
亮点:基于上述结论,我们提出了一种自关注 RoI 特征提取器( SAR FE ),以提高结构信息的特征提取的 3D 建议。
第十四条标题:抗噪:基于区域不确定性量化的半监督学习
作者:王振宇;李雅丽;郭业;王胜金
类别: cs.简历 [cs.简历|
亮点:在本文中,我们深入研究半监督学习的对象检测,标记的数据是更劳动密集型的收集。
第 15 条,标题:野外家庭识别( RFI W ):第 5 版
作者:约瑟夫 P 罗宾逊等。
类别: cs.简历 [cs.简历|
亮点:在本文中,我们总结了今年的 RFI W 的三个任务:具体来说,我们审查的结果亲属验证,三个主题验证,家庭成员的搜索和检索。
第十六条,标题:基于多种定价因素的枣果分类的机器学习方法
作者:阿布杜拉·扎凯里;鲁霍拉·赫达亚提;*·凯德马蒂;麦兰·塔吉布·戈尔吉科莱
类别: cs.简历 [cs.简历|
亮点:在本文中,我们提出了一种基于计算机视觉的方法,使用机器学习技术,将考虑到最重要的定价因素,可以用来增加农民的利润枣果分级。
第 17 条,标题:面具和无面具在儿童人脸识别中的纵向分析
作者:普拉文·库马尔·钱达利亚;扎希德·阿赫塔尔;尼塔·纳因
类别: cs.简历 [cs.简历|
亮点:因此,本研究的主要目的是分析儿童的纵向影响与面罩和其他协变量的人脸识别系统。
第 18 条,标题:卷积神经网络的潜遍历可视化解释
作者:阿米尔·德拉维德;艾格洛斯·卡萨格洛斯
分类:综合类【综合类】cs.LG,[见第四章,第一. 5 . 4 节;第一. 5 . 1 节、第一. 4 . 9 节和第一.2.10节]
亮点:使用 COVID - 19 胸部 X 光片,我们提出了一种方法来解释什么 CNN 已经利用生成对抗网络( GANs )。
第 19 条标题:学习迭代鲁棒变换同步
作者:紫健耀;李金熙
类别:综合类
亮点:在这项工作中,我们避免手工强大的损失函数,并建议使用图神经网络(GNNs)要学会转化同步。
第 20 条,标题:三种方法促进DNN对分布外方向和照明中对象的泛化:延迟停止、调整批规范化和不变性损失
作者:酒井明等。
类别: cs.简历 [cs.简历, cs.艾, cs.LG |
亮点:在本文中,我们调查了三种不同的方法,以改善DNNs在 OOD 方向和光照识别物体。
第 21 条,标题:学习图像压缩与分离的超先验解码器
作者:赵赞;刘超;孙鹤鸣;曾晓阳;范一波
类别:综合类【综合类】
亮点:在本文中,我们发现,性能瓶颈在于使用一个单一的超先验解码器,在这种情况下,三进制高斯模型崩溃到一个二进制的。
第 22 条,标题:智能(采样)增强:语义分割的最优有效数据增强
作者:米斯加纳内加西;黛安瓦格纳;亚历山大赖特尔
类别: cs.简历 [cs.简历, cs.LG |
亮点:现有的工作主要集中在图像分类和对象检测,而我们提供的语义图像分割的第一个研究,并引入两个新的方法:\ text it { Smart Automation }和\textit{SmartSamplingAugment}.
第二十三条,标题:基于局部三维深度描述符的环闭合检测
作者:周有杰;王一鸣;法比奥波耶西;秦齐;万一
类别:综合类【综合类】
亮点:我们提出了一个简单而有效的方法来解决环闭合检测同时本地化和映射使用本地 3D 深度描述符( L3 Ds )。
第 24 条,标题:基于局部突触可塑性学习事件的时空特征描述子:计算机视觉的生物现实视角
作者:沙法等。
类别: cs.简历 [cs.简历, cs.NE]
亮点:我们提出了一个基于优化的理论描述扣球皮层集合装备的 Spike - Timing - Dependent 可塑性(STDP)学习,如经验性地观察到的视觉皮层。
第 25 条,标题:3DP3:基于概率规划的三维场景感知
作者:尼沙德·戈索斯卡等。
类别: cs.简历 [cs.简历, cs.AI|
亮点:我们提出了3DP3,一个框架的逆向图形,使用推理在一个结构化的生成模型的对象,场景和图像。
第 26 条标题:具有凸形状先验性的测地模型
作者:陈达;米雷博;舒明磊;戴学成;科恩
类别: cs.简历 [cs.简历|
亮点:在本文中,我们考虑到一个更复杂的问题:找到曲率惩罚测地线路径与凸形状的先验。
第 27 条,标题:基于开放世界抽样的不平衡种子数据对比学习改进
作者:姜子瑜;陈天龙;陈婷;王章洋
类别: cs.简历 [cs.简历|
亮点:在这项工作中,我们提出了一个开放世界的未标记的数据采样框架称为模型-预警 K -中心( MAK ),它遵循三个简单的原则:( 1 )尾性,鼓励抽样的例子,从尾类,通过排序的经验对比损失期望( EC LE )的随机数据扩增的样本;( 2 )接近,拒绝的分布外的离群值,可能会分散训练;( 3 )多样性,确保多样性的采样的例子。
第 28 条,标题:HIER MATCH :利用标签层次提高半监督学习
作者:阿诗玛·加尔;肖里亚·巴加;亚什瓦尔丹·辛格;萨凯·阿南
类别: cs.简历 [cs.简历|
亮点:朝着提高性能的半监督学习方法的目标,我们提出了一个新的框架, HIER MATCH ,半监督的方法,利用分层信息,以减少标签的成本,并执行以及香草半监管学习方法。
第 29 条,标题:规划的 GANs 收敛更快
作者:阿克塞尔·索尔;卡什亚普·奇塔;延斯·穆勒;安德烈亚斯·盖格尔
类别: cs.简历 [cs.简历, cs.LG |
亮点:动机的发现,歧视者不能充分利用从更深层次的预训练模型的功能,我们提出了一个更有效的策略,混合通道和分辨率的功能。
第 30 条,标题:基于 SIFT 特征的虹膜识别
作者:费尔南多·阿隆索-费尔南德斯;佩德罗·汤姆-冈萨雷斯;弗吉尼亚·鲁伊斯-阿尔巴赛特;哈维尔·奥尔特加-加西亚
类别: cs.简历 [cs.简历|
亮点:在本文中,我们使用的尺度不变特征变换( SIFT )识别虹膜图像。
第 31 条,标题:在我的时间背景的一点帮助下:多模态自我中心动作识别
作者:哈札科斯;许哲成;纳格拉尼;齐泽曼;达曼
类别:综合类【综合类】cs.SD,易斯]
亮点:我们利用行动的时间背景,并提出了一种方法,学习到周围的行动,以提高识别性能。
第 32 条,标题:虹膜识别中的伪图像直接攻击
作者:弗吉尼亚·鲁伊斯-阿尔巴赛特等。
类别: cs.简历 [cs.简历|
亮点:在这方面的贡献,基于虹膜识别系统的直接攻击的脆弱性进行了研究。
第 33 条,标题:面向健壮问答的自省式蒸馏
作者:牛玉磊;张汉旺
类别: cs.简历 [cs.简历, cs.艾, cs.CL|
亮点:在本文中,我们提出了一种新的去偏方法称为内省蒸馏( Intro D ),使两个世界的最佳 QA 。
第 34 条,标题:有量词和无量词 CG ANs 的统一观点
作者:陈思安;李春亮;林宣天
类别: cs.简历 [cs.简历, cs.LG |
亮点:在这项工作中,我们证明了分类器可以适当地利用,以提高 cG ANs 。
第 35 条,标题:DIB - R ++:学习用混合可微渲染器预测光线和材料
作者:陈文正等。
类别:综合类【综合类】cs.GR]
亮点:在这项工作中,我们提出了 DI BR ++,混合可微渲染器,支持这些照片的真实效果相结合的光栅化和射线跟踪,利用各自的优势-速度和现实主义。
第 36 条,标题:基于高斯核混合网络的单幅图像离焦去模糊
作者:全玉辉;吴子聪;季慧
类别: cs.简历 [cs.简历|
亮点:本文提出了一种端到端的深度学习方法来消除散焦模糊从一个单一的图像,从而有一个所有的焦点图像,随后的视觉任务。
第 37 条,标题:语义分割的深度确定性不确定性
作者:吉什努·穆克蒂;约斯特·范阿姆斯福特;菲利普·托尔;亚林·加尔
类别: cs.简历 [cs.简历, cs.LG |
亮点:我们扩展的深度确定性的不确定性( DDU ),使用特征空间密度的方法进行不确定估计,语义分割。
第 38 条,标题:智能手机自注意移动网络图像倾斜校正的简单方法
作者:西丹·加格;黛比·普拉桑娜·莫汉蒂;湿婆·普拉萨德·托塔;苏库马尔·莫哈拉纳
类别: cs.简历 [cs.简历, cs.AI|
亮点:我们工作的主要贡献有两方面。
第 39 条,标题:用变压器监测牲畜
作者: Bhavesh Tangir ala 等。
类别: cs.简历 [cs.简历, cs.AI|
亮点:我们提出了 star former ,第一个端到端的多对象牲畜监测框架,学习实例级嵌入分组猪通过使用变压器架构。
第 40 条,标题:设备上实时手势识别
作者:宋志强等。
类别: cs.简历 [cs.简历|
亮点:我们提出了一个在设备上的实时手势识别(HGR)系统,该系统从单个 RGB 摄像头检测一组预定义的静态手势。
第 41 条,标题:CvS :小数据集的分割分类
作者:诺辛莫加卜;菲利普余;乔艾尔哈哈拉克;达文易
类别: cs.简历 [cs.简历|
亮点:本文提出了 CvS ,一个具有成本效益的分类器的小数据集,从预测的分割地图的分类标签。
第 42 条,标题:符号语言理解模式:以尼日利亚手语为例
作者:史蒂文科拉沃勒;奥佩米奥萨库德;纳扬萨克森娜;巴巴通德卡齐姆奥洛里萨德
类别: cs.简历 [cs.简历|
亮点:通过本文,我们试图减少听力障碍的社区和更大的社会之间的沟通障碍,通常不熟悉手语在非洲撒哈拉以南地区的听力残疾案件发生率最高,同时使用尼日利亚作为案例研究。
第 43 条,标题:基于可穿戴相机和多模态融合的以自我为中心的人体轨迹预测
作者:邱佳宁等。
类别: cs.简历 [cs.简历|
亮点:在本文中,我们解决的问题,预测的轨迹的自我为中心的相机佩戴者(自我的人)在拥挤的空间。
第 44 条,标题:DRB ANET :一种带边界辅助的轻量级语义分割双分辨率网络
作者:王林杰;周泉;姜晨峰;吴晓富;拉泰基
类别: cs.简历 [cs.简历|
亮点:本文介绍了一种轻量级的双分辨率网络,称为 DRB ANet ,旨在改善语义分割结果的援助边界信息。
第 45 条,标题:对比学习在什么情况下能保持从预训练到微调的对抗鲁棒性?
作者:范丽洁;刘思佳;陈品瑜;张高远;甘闯
类别: cs.简历 [cs.简历, cs.艾, cs.LG |
亮点:配备我们的新设计,我们提出了 Adv CL ,一种新的对抗对比的预训练框架。
第 46 条,标题:基于全卷积连体神经网络的卫星图像建筑物损伤评估
作者:叶甫盖尼;塔蒂阿娜
类别:综合类【综合类】
亮点:在这项工作中,我们开发了一种计算方法,用于自动比较同一地区的卫星图像在灾难之前和之后,并在建筑物中分类不同程度的损坏。
第四十七条标题:重新引导 AC GAN :具有稳定训练的辅助分类器 GAN
作者:姜明国;沈佑妍;赵敏素;朴载植
分类:综合类【综合类】cs.LG]
亮点:在本文中,我们介绍了两个治愈的 AC GAN 。
第四十八条,标题:基于 YOLO v4 深度神经网络的快速精确细粒级目标检测模型
作者:阿鲁纳哈· M ·罗伊;里奇·博斯;贾雅布拉塔·巴杜里
CATEGORY: cs.CV [cs.CV, cs.LG, 68T01, 68T05, 68T07, 68T10, 68T40, 68T45, 68U10,, I.4.9; I.5.2; I.5.4; I.2.1; I.2.9; I.2.m; J.7]
亮点:本文提出了一种高性能的实时细粒目标检测框架,解决了植物病害检测中的几个障碍,阻碍了传统方法的性能,如,密集分布,不规则的形态,多尺度对象类,纹理相似性等。
第 49 条,标题:流形上的地理感知分层贝叶斯学习
作者:范永辉;王亚林
类别: cs.简历 [cs.简历|
亮点:在本文中,我们提出了一个分层贝叶斯学习模型来解决这个挑战。
第 50 条,标题:Render in - between :用于动作插值的运动引导视频合成
作者:何宣一;徐晨;宋洁;希利日斯
类别: cs.简历 [cs.简历|
亮点:我们建议解决这些问题,在一个运动引导的帧上采样的框架,能够产生现实的人体运动和外观。
第 51 条,标题:语义分割中的去偏和去纠缠表示学习
作者:朱桑旭;金东万;韩伯雄
类别: cs.简历 [cs.简历, cs.LG |
亮点:为此,我们提出了一个模型不可知的和随机的训练计划,语义分割,有利于学习的去偏和不纠缠的表示。
第 52 条标题:腐败固定人重新识别的基准
作者:陈明辉;王志强;郑峰
类别: cs.简历 [cs.简历|
亮点:在这项工作中,我们全面建立六 Re ID 基准学习腐败不变的表示。
第 53 条标题:Face Scape :用于单视图三维人脸重建的三维人脸数据集和基准
作者:朱浩等。
类别:综合类【综合类】cs.GR]
亮点:在本文中,我们提出了一个大规模的详细的三维人脸数据集, Face Scape ,和相应的基准评估单视图面部三维重建。
第 54 条,标题:基于字面玩具数据集的分层图像分类
作者:贺龙;宋丹丹;梁政
类别: cs.简历 [cs.简历|
亮点:在本文中,我们提高了性能的分类融合功能,从层次结构的标签。
第 55 条,标题:基于碰撞感知因子的交互手的单目三维重建
作者:于荣;王静波;刘紫薇;陈变莱
类别: cs.简历 [cs.简历|
亮点:在本文中,我们作出了第一次尝试,重建 3D 交互手单目 RGB 图像。
第五十六届会议,标题:PA Net :具有动态接收域和自扩散监督的前瞻性感知网络
作者:陈小双等。
类别: cs.简历 [cs.简历|
亮点:在本文中,我们提出了一种新的角度感知的方法称为 PA Net 解决的角度问题。
第五十七条标题:基于偏倚矫正模块的改进局部表示的少镜头学习
作者:董超;叶琪;孟文超;杨开祥
类别: cs.简历 [cs.简历|
亮点:在本文中,我们提出了一个深度偏置矫正网络( DB RN ),以充分利用存在于结构的特征表示的空间信息。
第 58 条,标题:注意特征聚集的稠密预测
作者:杨永旭;黄永东;包玉刚;彼得·孔斯基德;余旭
类别: cs.简历 [cs.简历|
亮点:在本文中,我们引入注意的特征聚合( AFA )融合不同的网络层与更表达的非线性操作。
第 59 条标题:PP - Shi Tu :一种实用的轻量级图像识别系统
作者:魏胜玉等。
类别: cs.简历 [cs.简历|
亮点:在本文中,我们提出了一个实用的轻量级图像识别系统,命名为 PP - Shi Tu ,包括以下三个模块,主体检测,特征提取和矢量搜索。
第 60 条,标题:Ada Pool :指数自适应池的信息保留下采样
作者:亚历山大·斯特久;罗纳德·波普
类别: cs.简历 [cs.简历|
亮点:为此,我们提出了一个自适应和指数加权池方法命名为$\ text it { ada Pool }$。
第 61 条,标题:LSTA - Net :面向骨骼动作识别的长、短期时空聚合网络
作者:陈太林;王世东;周德森;关玉
类别: cs.简历 [cs.简历|
亮点:为了解决这个问题,在这项工作中,我们提出了 LST A - Net :一个新的长期短期时空聚合网络,它可以有效地捕捉到长/短的依赖在一个时空的方式。
第 62 条,标题:多智能体感知的蒸馏协作图学习
作者:李一鸣等。
类别:综合类【综合类】
亮点:为了促进更好的性能带宽权衡多代理感知,我们提出了一种新的蒸馏协作图( Disco Graph )模型可训练,位置感知,和自适应的代理之间的协作。
第 63 条,标题:基于特征丰富度的目标检测器提取
作者:杜志兴等。
类别: cs.简历 [cs.简历|
亮点:为了解决上述问题,我们提出了一种新的特征-粗糙度得分( FRS )的方法来选择重要的功能,提高广义检测提取过程中。
第 64 条,标题:具有鲁棒最优传输的精确点云配准
作者:沈正阳等。
类别:综合类【综合类】cs.CG,一.2.10
亮点:这项工作研究了使用鲁棒的最优传输( OT )的形状匹配。
第 65 条,标题:使用一种新的独特的人的外貌数据集的人和机器的人脸检测的评估
作者:内杰特·古尔坎
类别:综合类
亮点:为了实现这一目标,我们收集了独特的人的外观数据集,图像集,表示外观与低频,往往是采样不足的人脸数据集。
第 66 条,标题:Tri VoC :具有极端离群率的鲁棒点云配准的基于投票的有效共识最大化
作者:孙雷;邓路
类别:综合类【综合类】
亮点:在本文中,我们提出了一种新的,快速,确定性和保证的鲁棒求解器,命名为 Tri VoC (三层投票与共识最大化),鲁棒注册问题。
第 67 条,标题:RM Net :等效地去除网络中的剩余连接
作者:孟凡旭;郝成;庄家新;李克;孙兴
类别: cs.简历 [cs.简历, cs.LG |
亮点:在本文中,我们的目的是补救这个问题,并提出删除的保留和合并( RM )操作在 ResBlock 香草 ResNet 的等价连接。
第 68 条,标题:模仿任意说话风格的真实音频驱动说话人脸合成
作者:吴昊哲等。
类别:综合类【综合类】cs.GR,cs.MM,第一.1.4段
亮点:在本文中,我们建议注入风格的谈话脸合成框架,通过模仿任意说话风格的特定参考视频。
第 69 条,标题:近年来几种稀疏目标检测算法的比较研究
作者:冷家旭等。
类别: cs.简历 [cs.简历, cs.AI|
亮点:按照这种分类,我们提出了一个正式的定义,主要的挑战,基准数据集,评估指标,学习策略的重要审查。
第 70 条,标题:用于二维地形标志检测的特征聚集和细化网络
作者:敖跃元;吴洪
类别: cs.简历 [cs.简历|
亮点:在本文中,我们提出了一种新的深度网络,称为特征聚合和细化网络( FAR Net ),用于自动检测解剖标志。
第 71 条,标题:Patch Former :一种基于贴片注意的通用 3D 变换器
作者:张成;万浩成;沈欣怡;吴子钊
类别:综合类
亮点:为了解决这个缺点,我们引入补丁注意自适应学习一个更小的基础上计算的注意地图集。
第 72 条,标题:两个脑袋总比一个好:点云分类和分割的几何-潜伏注意
作者:韦拉斯奎兹;加列戈;费希尔
类别:综合类【综合类】cs.GR,cs.LG]
亮点:我们提出了一个创新的双头的注意层,结合几何和潜在的功能分割成语义有意义的子集的 3D 场景。
第 73 条,标题:MF Net :基于像素度量学习的多类少镜头分割网络
作者:张淼;史妙景;李丽
类别: cs.简历 [cs.简历|
亮点:这项工作的重点是少镜头语义分割,这在很大程度上仍然是一个未开发的领域。
第 74 条,标题:一种用于电影人物行为实例搜索的时空身份验证方法
作者:杨静瑶;梁超;牛彦瑞;黄保金;王中原
类别:彩色电视【彩色电视,红外电视,cs.MM]
亮点:具体而言,在空间维度上,我们提出了一个身份一致性验证方案来优化人 INS 和动作 INS 的直接融合得分。
第 75 条,标题:MOST - GAN :用于分离人脸图像处理的 3D 可变形样式 GAN
作者:萨法·梅丁等。
分类:综合类【综合类】cs.GR,cs.LG,一.2.10
亮点:相比之下,我们提出了一个框架,先验模型的物理属性的脸,如 3D 形状,反照率,姿势,照明明确,从而提供解纠缠的设计。
第 76 条,标题:利用 SE ( 3 )等方差自监控类别级目标姿态估计
作者:李小龙等。
类别: cs.简历 [cs.简历|
亮点:为了减少类别级学习所需的大量的姿势注释,我们首次提出了一个自我监督的学习框架,估计类别级的 6D 对象从单一的 3D 点云构成。在训练过程中,我们的方法假设没有地面真实的姿势注释,没有 CAD 模型,没有多视图监督。
第 77 条,标题:FREGAN :生成对抗网络在提高视频帧率中的应用
作者:里希克·米什拉;尼拉吉·古普塔;尼蒂亚·舒克拉
类别:综合类【综合类】cs.LG,第四章,第一.2.1节]
亮点:在本文中,我们研究了 GAN 模型,并提出了 FR EGAN 视频帧速率的提高。
第 78 条,标题:Oct Field :用于三维建模的分层隐函数
作者:唐继亨等。
类别:cs.GR[cs.GR,国别方案】
亮点:在这项工作中,我们提出了一个可学习的分层隐式表示的三维表面,编码 Oct Field ,允许高精度编码的复杂的表面与低内存和计算预算。
第 79 条,标题:单品时尚推荐:迈向跨领域推荐
作者:赛义德·奥米德·穆罕默迪;侯赛因·博达吉;艾哈迈德·卡尔霍尔
类别:中红外【中红外,中紫外】cs.LG]
亮点:尽管如此,许多挑战还在前面,本研究试图解决一些。
第 80 条,标题:具有注意机制的嵌套多实例学习
作者:索尔·福斯特;特里格夫·埃夫斯特·勒;基尔斯蒂·恩根
类别:cs.LG[cs.LG,国别方案】
亮点:在经典的图像数据集的实验表明,我们提出的模型提供了高精度的性能,以及在图像区域上发现相关的实例。
第 81 条,标题:FC 2T2 :快速连续卷积 Taylor 变换及其在视觉和图形中的应用
作者:海宁·兰格;内森·库兹
类别:cs.LG[cs.LG,行政长官,行政长官]
亮点:在本文中,我们从现代机器学习的角度重新审视泰勒级数展开。
第 82 条,标题:预测大西洋的年代际变化
作者:刘辉;王培东;马修贝弗里奇;权英王;伊多德罗里
类别:cs.LG[cs.LG,中文版,物理版
亮点:我们使用的数据来自社区地球系统模型 1 大集合项目,一个国家的最先进的气候模型与 3440 年的数据。
第 83 段,标题:被愚弄的正确原因:通过教师指导的课程学习方法提高对抗鲁棒性
作者:阿尼迪亚·萨卡;阿尼班·萨卡;索里亚·加利;维内斯· N ·巴拉苏布拉马尼安
类别: cs.LG [cs.LG, cs.Cr, cs.简历|
亮点:我们提出了一个非迭代的方法,在训练中实施以下想法。
第 84 条,标题:关于比较式自我监控学习的推广
作者:黄蔚然;易明阳;赵旭阳
类别:cs.LG[cs.LG,[美、英、美、日、美]
亮点:为此,我们提出了一个理论解释的对比自我监督预训练模型推广到下游任务。
第 85 条,标题:用有限的数据掌握雅达利游戏
作者:叶伟瑞;刘少怀;库鲁塔奇;彼得·阿贝尔;杨高
类别:cs.LG[cs.LG,[行政长官、副长官、行政长官]
亮点:我们提出了一个示例有效的基于模型的可视化 RL 算法建立在 Mu Zero 上,我们命名为 Efficient Zero 。
第 86 段,标题:伪反转编码器
作者:詹杰泽贝特勒;伊万索斯诺维奇;阿诺德斯穆德斯
类别: cs.LG [cs.LG, cs.简历|
亮点:我们介绍了一类新的基于似然的伪双射结构的自动编码器,我们称之为伪可逆编码器。
第 87 条,标题:广播体育视频理解中的远程监督语义文本检测与识别
作者:阿维吉特·沙;托普乔伊·比斯瓦斯;萨西什·拉马多斯;德文·桑托什·沙
类别:cs.MM[cs.MM,CV 系统和 IR 系统cs.LG,第一条第 1 款;第一. 4 . 8 条;第二. 10 条;第一. 2 . 7 条和第一. 4 . 9 条)
亮点:在这项工作中,我们研究了非常精确的语义文本检测和识别运动时钟,并提出挑战。最后,我们分享我们的计算架构管道,以规模在工业环境中的系统,并提出了一个强大的数据集同样验证我们的结果。
第 88 条,标题:时间矩局部化的分层深度残差推理
作者:马*;韩先静;宋雪萌;崔怡然;聂立强
类别:cs.MM[cs.MM,cs.CL,[英属维尔京群岛]
亮点:为此,我们提出了一个层次的深剩余推理(HDRR)模型,该模型将视频和句子分解成具有不同语义的多级表示,以实现更细粒度的本地化。
第 89 条,标题:焦点注意网络:生物医学图像分割中的注意优化
作者:杨德昌;杨光;艾维丝萨拉;卡罗拉-比比安-施奈利布;
类别: ees. iv [ees. iv, cs.艾, cs.简历, 数学. Oc |
亮点:在本文中,我们调查的作用,调节注意的焦点参数,揭示了注意之间的联系,在损失函数和网络。
第九十届会议,标题:使用 Grad CAM 模拟真实的 MRI 变化以改进深度学习模型和可视化解释
作者:*·伊利亚斯·帕特尔等。
类别:第四组[第四、第六、第五组]
亮点:我们使用一种改进的 High Res 3 DNet 模型来解决脑 MRI 体积地标检测问题。
第 91 条,标题:To rch XRay Vision :胸部 X 线数据集和模型库
作者:柯亨等。
类别:第四组[第四、第六、第五组]
亮点: TorchX Ray Vision :胸透数据集和模型库
第 92 条,标题:I GCN :用于 2D / 3D 变形配准的图像到图卷积网络
作者:中尾惠美;中村光弘;松田哲也
类别: EESS . IV 【 ESS . IV ,第四章,第五章】cs.LG]
亮点:我们提出了一个图像到图形的卷积网络,实现变形配准的 3D 器官网格的单视点 2D 投影图像。
第 93 条,标题:三维脑肿瘤 MRI 语义分割中的冗余度减少
作者:马福祖尔·拉赫曼·西迪基;安德里亚·米罗年科
类别:第四组【第四、第五组】
亮点:在这项工作中,我们保持了基于编码解码器的分割网络,但集中在网络训练过程,最大限度地减少冗余扰动的修改。
第九十四条,标题:函数神经网络在参数图像复原中的应用
作者:罗芳舟;吴晓林;郭艳辉
类别:第四组【第四、第五组】
亮点:在这项工作中,我们提出了一种新的系统称为功能神经网络( Fun cNet ),以解决一个单一的模型的参数图像恢复问题。
第 95 号决议,标题:M2 MRF :用于眼底图像微小病灶分割的多对多特征重组
作者:刘庆;刘昊天;梁一雄
类别:第四组【第四、第五组】
亮点:在本文中,我们提出了一个多对多功能重组( M2 MRF )。
第九十六条,标题:稳健的深度生成先验重建和差分配准胎儿 MRI 在孕龄预测中的应用
作者:卢西里奥·科德罗·格兰代等。
类别:eess.IV,cs.CV,物理学。med-ph,92C50,94A08,I.2.1;J.3]
亮点:因此,在本文中,我们提出了一个强大的体积重建集成的微分同胚体积切片注册方法的深度生成前。
第 97 条,标题:用于心率和呼吸频率估计的双注意网络
作者:任玉琢;布赖登·瑟尔尼克;尼兰詹·阿瓦达南
==同步,由老年人纠正=@EERDER_MAN
亮点:我们提出了一个卷积神经网络,利用空间注意和渠道注意,我们称之为双注意网络( DAN ),共同估计心率和呼吸频率与摄像机视频作为输入。
第 98 条,标题:用于多光谱目标检测的互模融合变压器
作者:方青云;韩大鹏;王兆奎
类别:第四组【第四、第五组】
亮点:为了充分利用不同的模式,我们提出了一个简单而有效的跨模态特征融合方法,称为跨模态融合变压器(CFT)在本文中。
第九十九条,标题:将边界不确定性引入损失函数的生物医学图像分割
作者:杨广;杨光;艾维丝萨拉;卡罗拉-比比安;列奥纳多朗多
类别: ees. iv [ees. iv, cs.艾, cs.简历, 数学. Oc |
亮点:在本文中,我们提出了边界的不确定性,它使用形态学操作,限制软标签的对象边界,提供了一个适当的表示地面的真相标签的不稳定性,并可以适应,使强大的模型训练系统手动分割错误的存在。
第 100 条,标题:生物医学图像分割中的神经网络过置信度校正
作者:杨明等。
类别: ees. iv [ees. iv, cs.艾, cs.简历, 数学. Oc |
亮点:在这项研究中,我们确定差校准作为基于深度学习的生物医学图像分割的一个新兴挑战。
第 101 条,标题:磁共振图像质量指标与神经网络分割精度的相关性
作者:拉贾耶斯瓦里·穆斯蒂瓦拉詹等。
类别:第四组【第四、第五组】
磁共振图像质量指标与神经网络分割精度的相关性
第 102 段,标题:图像去噪中的自确认
作者:林黄兴等。
类别:第四组【第四、第五组】
亮点:我们设计了一种新的正则化,称为自验证,图像去噪。
第 103 条,标题:少数镜头学习的有影响力的原型网络:皮肤病案例研究
作者:兰加纳·罗伊·乔杜里;迪普提· R ·巴苏拉
类别: EESS . IV 【 ESS . IV ,第四章,第五章】cs.LG]
亮点:在这项工作中,我们提出了一个新的版本的 PN 属性的权重,以支持样本相应的支持样本分布的影响。
第 104 条,标题:高动态范围图像色调映射的非成对学习
作者:雅埃尔·温克;因巴尔·休伯曼-斯皮格尔格拉斯;拉南·法塔尔
类别: EESS . IV 【 ESS . IV ,第四章,第五章】cs.LG]
亮点:在本文中,我们描述了一种新的色调映射方法的指导下,产生低动态范围的明确目标(LDR)最好地再现了原生的视觉特征的渲染LDR图像
第 105 号,标题:用 77 GHz 雷达学习探测明火和隐蔽物
作者:高翔宇等。
类别:斯佩斯。Sp, cs.简历|
亮点:在本文中,我们专注于使用低成本的 77 GHz 毫米波雷达携带的目标检测问题的相对未开发的领域。
1, TITLE: Robustness of Deep Learning Algorithms in Astronomy -- Galaxy Morphology Studies
AUTHORS: A. ?IPRIJANOVI? et. al.
CATEGORY: astro-ph.GA [astro-ph.GA, cs.CV, cs.LG]
HIGHLIGHT: Deep learning models are being increasingly adopted in wide array of scientific domains, especially to handle high-dimensionality and volume of the scientific data.
2, TITLE: A Robust Single-pixel Particle Image Velocimetry Based on Fully Convolutional Networks with Cross-correlation Embedded
AUTHORS: QI GAO et. al.
CATEGORY: physics.flu-dyn [physics.flu-dyn, cs.AI, cs.CV]
HIGHLIGHT: In the current work, we propose a new velocity field estimation paradigm, which achieves a synergetic combination of the deep learning method and the traditional cross-correlation method.
3, TITLE: DeepDoseNet: A Deep Learning Model for 3D Dose Prediction in Radiation Therapy
AUTHORS: Mumtaz Hussain Soomro ; Victor Gabriel Leandro Alves ; Hamidreza Nourzadeh ; Jeffrey V. Siebers
CATEGORY: physics.med-ph [physics.med-ph, cs.AI, cs.CV, cs.LG, eess.IV]
HIGHLIGHT: The DeepDoseNet 3D dose prediction model based on ResNet and Dilated DenseNet is proposed.
4, TITLE: From Face to Gait: Weakly-Supervised Learning of Gender Information from Walking Patterns
AUTHORS: Andy Catruna ; Adrian Cosma ; Ion Emilian Radoi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We propose a weakly-supervised method for learning gender information of people based on their manner of walking.
5, TITLE: Higher-Order Implicit Fairing Networks for 3D Human Pose Estimation
AUTHORS: Jianning Quan ; A. Ben Hamza
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we introduce a higher-order graph convolutional framework with initial residual connections for 2D-to-3D pose estimation.
6, TITLE: Generalized Data Weighting Via Class-level Gradient Manipulation
AUTHORS: CAN CHEN et. al.
CATEGORY: cs.CV [cs.CV, cs.IT, cs.LG, math.IT]
HIGHLIGHT: To this end, in this paper, we propose Generalized Data Weighting (GDW) to simultaneously mitigate label noise and class imbalance by manipulating gradients at the class level.
7, TITLE: Polyline Based Generative Navigable Space Segmentation for Autonomous Visual Navigation
AUTHORS: Zheng Chen ; Zhengming Ding ; David Crandall ; Lantao Liu
CATEGORY: cs.CV [cs.CV, cs.RO]
HIGHLIGHT: In this work, we treat the visual navigable space segmentation as a scene decomposition problem and propose Polyline Segmentation Variational AutoEncoder Networks (PSV-Nets), a representation-learning-based framework to enable robots to learn the navigable space segmentation in an unsupervised manner.
8, TITLE: VPFNet: Voxel-Pixel Fusion Network for Multi-class 3D Object Detection
AUTHORS: Chia-Hung Wang ; Hsueh-Wei Chen ; Li-Chen Fu
CATEGORY: cs.CV [cs.CV, cs.AI, cs.LG, cs.RO]
HIGHLIGHT: In order to elevate the detection performance in a complicated environment, this paper proposes a deep learning (DL)-embedded fusion-based multi-class 3D object detection network which admits both LiDAR and camera sensor data streams, named Voxel-Pixel Fusion Network (VPFNet).
9, TITLE: Generative Occupancy Fields for 3D Surface-Aware Image Synthesis
AUTHORS: Xudong Xu ; Xingang Pan ; Dahua Lin ; Bo Dai
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: In this paper, we propose Generative Occupancy Fields (GOF), a novel model based on generative radiance fields that can learn compact object surfaces without impeding its training convergence.
10, TITLE: DFCANet: Dense Feature Calibration-Attention Guided Network for Cross Domain Iris Presentation Attack Detection
AUTHORS: Gaurav Jaswal ; Aman Verma ; Sumantra Dutta Roy ; Raghavendra Ramachandra
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To alleviate these shortcomings, this paper proposes DFCANet: Dense Feature Calibration and Attention Guided Network which calibrates the locally spread iris patterns with the globally located ones.
11, TITLE: PP-PicoDet: A Better Real-Time Object Detector on Mobile Devices
AUTHORS: GUANGHUA YU et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this work, we are dedicated to studying key optimizations and neural network architecture choices for object detection to improve accuracy and efficiency.
12, TITLE: DPNET: Dual-Path Network for Efficient Object Detectioj with Lightweight Self-Attention
AUTHORS: * Shi ; Quan Zhou ; Yinghao Ni ; Xiaofu Wu ; Longin Jan Latecki
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To address the trade-off between computational cost and detection accuracy, this paper presents a dual path network, named DPNet, for efficient object detection with lightweight self-attention.
13, TITLE: Structure Information Is The Key: Self-Attention RoI Feature Extractor in 3D Object Detection
AUTHORS: Diankun Zhang ; Zhijie Zheng ; Xueting Bi ; Xiaojun Liu
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Based on the above conclusion, we propose a Self-Attention RoI Feature Extractor (SARFE) to enhance structural information of the feature extracted from 3D proposals.
14, TITLE: Combating Noise: Semi-supervised Learning By Region Uncertainty Quantification
AUTHORS: Zhenyu Wang ; Yali Li ; Ye Guo ; Shengjin Wang
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we delve into semi-supervised learning for object detection, where labeled data are more labor-intensive to collect.
15, TITLE: Recognizing Families In The Wild (RFIW): The 5th Edition
AUTHORS: JOSEPH P. ROBINSON et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we summarize submissions for the three tasks of this year's RFIW: specifically, we review the results for kinship verification, tri-subject verification, and family member search and retrieval.
16, TITLE: Classification of Jujube Fruit Based on Several Pricing Factors Using Machine Learning Methods
AUTHORS: Abdollah Zakeri ; Ruhollah Hedayati ; Mohammad Khedmati ; Mehran Taghipour-Gorjikolaie
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper we are proposing a computer vision-based method for grading jujube fruits using machine learning techniques which will take most of the important pricing factors into account and can be used to increase the profit of farmers.
17, TITLE: Longitudinal Analysis of Mask and No-Mask on Child Face Recognition
AUTHORS: Praveen Kumar Chandaliya ; Zahid Akhtar ; Neeta Nain
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Thus, the main objective of this study is analyzing the child longitudinal impact together with face mask and other covariates on face recognition systems.
18, TITLE: Visual Explanations for Convolutional Neural Networks Via Latent Traversal
AUTHORS: Amil Dravid ; Aggelos K. Katsaggelos
CATEGORY: cs.CV [cs.CV, cs.AI, cs.LG, eess.IV, I.5.4; I.5.1; I.4.9; I.2.10]
HIGHLIGHT: Using COVID-19 chest X-rays, we present a method for interpreting what a CNN has learned by utilizing Generative Adversarial Networks (GANs).
19, TITLE: Learning Iterative Robust Transformation Synchronization
AUTHORS: Zi Jian Yew ; Gim Hee Lee
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this work, we avoid handcrafting robust loss functions, and propose to use graph neural networks (GNNs) to learn transformation synchronization.
20, TITLE: Three Approaches to Facilitate DNN Generalization to Objects in Out-of-distribution Orientations and Illuminations: Late-stopping, Tuning Batch Normalization and Invariance Loss
AUTHORS: AKIRA SAKAI et. al.
CATEGORY: cs.CV [cs.CV, cs.AI, cs.LG]
HIGHLIGHT: In this paper, we investigate three different approaches to improve DNNs in recognizing objects in OoD orientations and illuminations.
21, TITLE: Learned Image Compression with Separate Hyperprior Decoders
AUTHORS: Zhao Zan ; Chao Liu ; Heming Sun ; Xiaoyang Zeng ; Yibo Fan
CATEGORY: cs.CV [cs.CV, eess.IV]
HIGHLIGHT: In this paper, we find that the performance bottleneck lies in the use of a single hyperprior decoder, in which case the ternary Gaussian model collapses to a binary one.
22, TITLE: Smart(Sampling)Augment: Optimal and Efficient Data Augmentation for Semantic Segmentation
AUTHORS: Misgana Negassi ; Diane Wagner ; Alexander Reiterer
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: Existing work focuses on image classification and object detection, whereas we provide the first study on semantic image segmentation and introduce two new approaches: \textit{SmartAugment} and \textit{SmartSamplingAugment}.
23, TITLE: Loop Closure Detection Using Local 3D Deep Descriptors
AUTHORS: Youjie Zhou ; Yiming Wang ; Fabio Poiesi ; Qi Qin ; Yi Wan
CATEGORY: cs.CV [cs.CV, cs.RO]
HIGHLIGHT: We present a simple yet effective method to address loop closure detection in simultaneous localisation and mapping using local 3D deep descriptors (L3Ds).
24, TITLE: Learning Event-based Spatio-Temporal Feature Descriptors Via Local Synaptic Plasticity: A Biologically-realistic Perspective of Computer Vision
AUTHORS: ALI SAFA et. al.
CATEGORY: cs.CV [cs.CV, cs.NE]
HIGHLIGHT: We present an optimization-based theory describing spiking cortical ensembles equipped with Spike-Timing-Dependent Plasticity (STDP) learning, as empirically observed in the visual cortex.
25, TITLE: 3DP3: 3D Scene Perception Via Probabilistic Programming
AUTHORS: NISHAD GOTHOSKAR et. al.
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: We present 3DP3, a framework for inverse graphics that uses inference in a structured generative model of objects, scenes, and images.
26, TITLE: Geodesic Models with Convexity Shape Prior
AUTHORS: Da Chen ; Jean-Marie Mirebeau ; Minglei Shu ; Xuecheng Tai ; Laurent D. Cohen
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we take into account a more complicated problem: finding curvature-penalized geodesic paths with a convexity shape prior.
27, TITLE: Improving Contrastive Learning on Imbalanced Seed Data Via Open-World Sampling
AUTHORS: Ziyu Jiang ; Tianlong Chen ; Ting Chen ; Zhangyang Wang
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this work, we present an open-world unlabeled data sampling framework called Model-Aware K-center (MAK), which follows three simple principles: (1) tailness, which encourages sampling of examples from tail classes, by sorting the empirical contrastive loss expectation (ECLE) of samples over random data augmentations; (2) proximity, which rejects the out-of-distribution outliers that may distract training; and (3) diversity, which ensures diversity in the set of sampled examples.
28, TITLE: HIERMATCH: Leveraging Label Hierarchies for Improving Semi-Supervised Learning
AUTHORS: Ashima Garg ; Shaurya Bagga ; Yashvardhan Singh ; Saket Anand
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Towards the goal of improving the performance of semi-supervised learning methods, we propose a novel framework, HIERMATCH, a semi-supervised approach that leverages hierarchical information to reduce labeling costs and performs as well as a vanilla semi-supervised learning method.
29, TITLE: Projected GANs Converge Faster
AUTHORS: Axel Sauer ; Kashyap Chitta ; Jens M�ller ; Andreas Geiger
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: Motivated by the finding that the discriminator cannot fully exploit features from deeper layers of the pretrained model, we propose a more effective strategy that mixes features across channels and resolutions.
30, TITLE: Iris Recognition Based on SIFT Features
AUTHORS: Fernando Alonso-Fernandez ; Pedro Tome-Gonzalez ; Virginia Ruiz-Albacete ; Javier Ortega-Garcia
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we use the Scale Invariant Feature Transformation (SIFT) for recognition using iris images.
31, TITLE: With A Little Help from My Temporal Context: Multimodal Egocentric Action Recognition
AUTHORS: Evangelos Kazakos ; Jaesung Huh ; Arsha Nagrani ; Andrew Zisserman ; Dima Damen
CATEGORY: cs.CV [cs.CV, cs.SD, eess.AS]
HIGHLIGHT: We capitalise on the action's temporal context and propose a method that learns to attend to surrounding actions in order to improve recognition performance.
32, TITLE: Direct Attacks Using Fake Images in Iris Verification
AUTHORS: VIRGINIA RUIZ-ALBACETE et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this contribution, the vulnerabilities of iris-based recognition systems to direct attacks are studied.
33, TITLE: Introspective Distillation for Robust Question Answering
AUTHORS: Yulei Niu ; Hanwang Zhang
CATEGORY: cs.CV [cs.CV, cs.AI, cs.CL]
HIGHLIGHT: In this paper, we present a novel debiasing method called Introspective Distillation (IntroD) to make the best of both worlds for QA.
34, TITLE: A Unified View of CGANs with and Without Classifiers
AUTHORS: Si-An Chen ; Chun-Liang Li ; Hsuan-Tien Lin
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: In this work, we demonstrate that classifiers can be properly leveraged to improve cGANs.
35, TITLE: DIB-R++: Learning to Predict Lighting and Material with A Hybrid Differentiable Renderer
AUTHORS: WENZHENG CHEN et. al.
CATEGORY: cs.CV [cs.CV, cs.GR]
HIGHLIGHT: In this work, we propose DIBR++, a hybrid differentiable renderer which supports these photorealistic effects by combining rasterization and ray-tracing, taking the advantage of their respective strengths -- speed and realism.
36, TITLE: Gaussian Kernel Mixture Network for Single Image Defocus Deblurring
AUTHORS: Yuhui Quan ; Zicong Wu ; Hui Ji
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This paper presents an end-to-end deep learning approach for removing defocus blur from a single image, so as to have an all-in-focus image for consequent vision tasks.
37, TITLE: Deep Deterministic Uncertainty for Semantic Segmentation
AUTHORS: Jishnu Mukhoti ; Joost van Amersfoort ; Philip H. S. Torr ; Yarin Gal
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: We extend Deep Deterministic Uncertainty (DDU), a method for uncertainty estimation using feature space densities, to semantic segmentation.
38, TITLE: A Simple Approach to Image Tilt Correction with Self-Attention MobileNet for Smartphones
AUTHORS: Siddhant Garg ; Debi Prasanna Mohanty ; Siva Prasad Thota ; Sukumar Moharana
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: The main contributions of our work are two-fold.
39, TITLE: Livestock Monitoring with Transformer
AUTHORS: BHAVESH TANGIRALA et. al.
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: We present starformer, the first end-to-end multiple-object livestock monitoring framework that learns instance-level embeddings for grouped pigs through the use of transformer architecture.
40, TITLE: On-device Real-time Hand Gesture Recognition
AUTHORS: GEORGE SUNG et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We present an on-device real-time hand gesture recognition (HGR) system, which detects a set of predefined static gestures from a single RGB camera.
41, TITLE: CvS: Classification Via Segmentation For Small Datasets
AUTHORS: Nooshin Mojab ; Philip S. Yu ; Joelle A. Hallak ; Darvin Yi
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This paper presents CvS, a cost-effective classifier for small datasets that derives the classification labels from predicting the segmentation maps.
42, TITLE: Sign-to-Speech Model for Sign Language Understanding: A Case Study of Nigerian Sign Language
AUTHORS: Steven Kolawole ; Opeyemi Osakuade ; Nayan Saxena ; Babatunde Kazeem Olorisade
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Through this paper, we seek to reduce the communication barrier between the hearing-impaired community and the larger society who are usually not familiar with sign language in the sub-Saharan region of Africa with the largest occurrences of hearing disability cases, while using Nigeria as a case study.
43, TITLE: Egocentric Human Trajectory Forecasting with A Wearable Camera and Multi-Modal Fusion
AUTHORS: JIANING QIU et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we address the problem of forecasting the trajectory of an egocentric camera wearer (ego-person) in crowded spaces.
44, TITLE: DRBANET: A Lightweight Dual-Resolution Network for Semantic Segmentation with Boundary Auxiliary
AUTHORS: Linjie Wang ; Quan Zhou ; Chenfeng Jiang ; Xiaofu Wu ; Longin Jan Latecki
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This paper introduces a lightweight dual-resolution network, called DRBANet, aiming to refine semantic segmentation results with the aid of boundary information.
45, TITLE: When Does Contrastive Learning Preserve Adversarial Robustness from Pretraining to Finetuning?
AUTHORS: Lijie Fan ; Sijia Liu ; Pin-Yu Chen ; Gaoyuan Zhang ; Chuang Gan
CATEGORY: cs.CV [cs.CV, cs.AI, cs.LG]
HIGHLIGHT: Equipped with our new designs, we propose AdvCL, a novel adversarial contrastive pretraining framework.
46, TITLE: Fully Convolutional Siamese Neural Networks for Buildings Damage Assessment from Satellite Images
AUTHORS: Eugene Khvedchenya ; Tatiana Gabruseva
CATEGORY: cs.CV [cs.CV, eess.IV]
HIGHLIGHT: In this work, we develop a computational approach for an automated comparison of the same region's satellite images before and after the disaster, and classify different levels of damage in buildings.
47, TITLE: Rebooting ACGAN: Auxiliary Classifier GANs with Stable Training
AUTHORS: Minguk Kang ; Woohyeon Shim ; Minsu Cho ; Jaesik Park
CATEGORY: cs.CV [cs.CV, cs.AI, cs.LG]
HIGHLIGHT: In this paper, we introduce two cures for ACGAN.
48, TITLE: A Fast Accurate Fine-grain Object Detection Model Based on YOLOv4 Deep Neural Network
AUTHORS: Arunabha M. Roy ; Rikhi Bose ; Jayabrata Bhaduri
CATEGORY: cs.CV [cs.CV, cs.LG, 68T01, 68T05, 68T07, 68T10, 68T40, 68T45, 68U10,, I.4.9; I.5.2; I.5.4; I.2.1; I.2.9; I.2.m; J.7]
HIGHLIGHT: This paper presents a high-performance real-time fine-grain object detection framework that addresses several obstacles in plant disease detection that hinder the performance of traditional methods, such as, dense distribution, irregular morphology, multi-scale object classes, textural similarity, etc.
49, TITLE: Geometry-Aware Hierarchical Bayesian Learning on Manifolds
AUTHORS: Yonghui Fan ; Yalin Wang
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a hierarchical Bayesian learning model to address this challenge.
50, TITLE: Render In-between: Motion Guided Video Synthesis for Action Interpolation
AUTHORS: Hsuan-I Ho ; Xu Chen ; Jie Song ; Otmar Hilliges
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: We propose to address these issues in a motion-guided frame-upsampling framework that is capable of producing realistic human motion and appearance.
51, TITLE: Learning Debiased and Disentangled Representations for Semantic Segmentation
AUTHORS: Sanghyeok Chu ; Dongwan Kim ; Bohyung Han
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: To this end, we propose a model-agnostic and stochastic training scheme for semantic segmentation, which facilitates the learning of debiased and disentangled representations.
52, TITLE: Benchmarks for Corruption Invariant Person Re-identification
AUTHORS: Minghui Chen ; Zhiqiang Wang ; Feng Zheng
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this work, we comprehensively establish six ReID benchmarks for learning corruption invariant representation.
53, TITLE: FaceScape: 3D Facial Dataset and Benchmark for Single-View 3D Face Reconstruction
AUTHORS: HAO ZHU et. al.
CATEGORY: cs.CV [cs.CV, cs.GR]
HIGHLIGHT: In this paper, we present a large-scale detailed 3D face dataset, FaceScape, and the corresponding benchmark to evaluate single-view facial 3D reconstruction.
54, TITLE: Hierarchical Image Classification with A Literally Toy Dataset
AUTHORS: Long He ; Dandan Song ; Liang Zheng
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we improve the performance of classification by fusing features learned from a hierarchy of labels.
55, TITLE: Monocular 3D Reconstruction of Interacting Hands Via Collision-Aware Factorized Refinements
AUTHORS: Yu Rong ; Jingbo Wang ; Ziwei Liu ; Chen Change Loy
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we make the first attempt to reconstruct 3D interacting hands from monocular single RGB images.
56, TITLE: PANet: Perspective-Aware Network with Dynamic Receptive Fields and Self-Distilling Supervision for Crowd Counting
AUTHORS: XIAOSHUANG CHEN et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a novel perspective-aware approach called PANet to address the perspective problem.
57, TITLE: Few-shot Learning with Improved Local Representations Via Bias Rectify Module
AUTHORS: Chao Dong ; Qi Ye ; Wenchao Meng ; Kaixiang Yang
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper we propose a Deep Bias Rectify Network (DBRN) to fully exploit the spatial information that exists in the structure of the feature representations.
58, TITLE: Dense Prediction with Attentive Feature Aggregation
AUTHORS: Yung-Hsu Yang ; Thomas E. Huang ; Samuel Rota Bul� ; Peter Kontschieder ; Fisher Yu
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we introduce Attentive Feature Aggregation (AFA) to fuse different network layers with more expressive non-linear operations.
59, TITLE: PP-ShiTu: A Practical Lightweight Image Recognition System
AUTHORS: SHENGYU WEI et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a practical lightweight image recognition system, named PP-ShiTu, consisting of the following 3 modules, mainbody detection, feature extraction and vector search.
60, TITLE: AdaPool: Exponential Adaptive Pooling for Information-Retaining Downsampling
AUTHORS: Alexandros Stergiou ; Ronald Poppe
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To this end, we propose an adaptive and exponentially weighted pooling method named $\textit{adaPool}$.
61, TITLE: LSTA-Net: Long Short-term Spatio-Temporal Aggregation Network for Skeleton-based Action Recognition
AUTHORS: Tailin Chen ; Shidong Wang ; Desen Zhou ; Yu Guan
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To address this issue, in this work we propose LSTA-Net: a novel Long short-term Spatio-Temporal Aggregation Network, which can effectively capture the long/short-range dependencies in a spatio-temporal manner.
62, TITLE: Learning Distilled Collaboration Graph for Multi-Agent Perception
AUTHORS: YIMING LI et. al.
CATEGORY: cs.CV [cs.CV, cs.RO]
HIGHLIGHT: To promote better performance-bandwidth trade-off for multi-agent perception, we propose a novel distilled collaboration graph (DiscoGraph) to model trainable, pose-aware, and adaptive collaboration among agents.
63, TITLE: Distilling Object Detectors with Feature Richness
AUTHORS: ZHIXING DU et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To address the above issues, we propose a novel Feature-Richness Score (FRS) method to choose important features that improve generalized detectability during distilling.
64, TITLE: Accurate Point Cloud Registration with Robust Optimal Transport
AUTHORS: ZHENGYANG SHEN et. al.
CATEGORY: cs.CV [cs.CV, cs.CG, I.2.10]
HIGHLIGHT: This work investigates the use of robust optimal transport (OT) for shape matching.
65, TITLE: Evaluation of Human and Machine Face Detection Using A Novel Distinctive Human Appearance Dataset
AUTHORS: Necdet Gurkan ; Jordan W. Suchow
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: Towards this goal, we collected the Distinctive Human Appearance dataset, an image set that represents appearances with low frequency and that tend to be undersampled in face datasets.
66, TITLE: TriVoC: Efficient Voting-based Consensus Maximization for Robust Point Cloud Registration with Extreme Outlier Ratios
AUTHORS: Lei Sun ; Lu Deng
CATEGORY: cs.CV [cs.CV, cs.RO]
HIGHLIGHT: In this paper, we present a novel, fast, deterministic and guaranteed robust solver, named TriVoC (Triple-layered Voting with Consensus maximization), for the robust registration problem.
67, TITLE: RMNet: Equivalently Removing Residual Connection from Networks
AUTHORS: Fanxu Meng ; Hao Cheng ; Jiaxin Zhuang ; Ke Li ; Xing Sun
CATEGORY: cs.CV [cs.CV, cs.LG]
HIGHLIGHT: In this paper, we aim to remedy this problem and propose to remove the residual connection in a vanilla ResNet equivalently by a reserving and merging (RM) operation on ResBlock.
68, TITLE: Imitating Arbitrary Talking Style for Realistic Audio-DrivenTalking Face Synthesis
AUTHORS: HAOZHE WU et. al.
CATEGORY: cs.CV [cs.CV, cs.GR, cs.MM, I.1.4]
HIGHLIGHT: In this paper, we propose to inject style into the talking face synthesis framework through imitating arbitrary talking style of the particular reference video.
69, TITLE: A Comparative Review of Recent Few-Shot Object Detection Algorithms
AUTHORS: LENG JIAXU et. al.
CATEGORY: cs.CV [cs.CV, cs.AI]
HIGHLIGHT: Following this taxonomy, we present a significant review of the formal definition, main challenges, benchmark datasets, evaluation metrics, and learning strategies.
70, TITLE: Feature Aggregation and Refinement Network for 2D AnatomicalLandmark Detection
AUTHORS: Yueyuan Ao ; Hong Wu
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: In this paper, we propose a novel deep network, named feature aggregation and refinement network (FARNet), for the automatic detection of anatomical landmarks.
71, TITLE: PatchFormer: A Versatile 3D Transformer Based on Patch Attention
AUTHORS: Zhang Cheng ; Haocheng Wan ; Xinyi Shen ; Zizhao Wu
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To solve this shortcoming, we introduce patch-attention to adaptively learn a much smaller set of bases upon which the attention maps are computed.
72, TITLE: Two Heads Are Better Than One: Geometric-Latent Attention for Point Cloud Classification and Segmentation
AUTHORS: Hanz Cuevas-Velasquez ; Antonio Javier Gallego ; Robert B. Fisher
CATEGORY: cs.CV [cs.CV, cs.GR, cs.LG]
HIGHLIGHT: We present an innovative two-headed attention layer that combines geometric and latent features to segment a 3D scene into semantically meaningful subsets.
73, TITLE: MFNet: Multi-class Few-shot Segmentation Network with Pixel-wise Metric Learning
AUTHORS: Miao Zhang ; Miaojing Shi ; Li Li
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: This work focuses on few-shot semantic segmentation, which is still a largely unexplored field.
74, TITLE: A Spatio-Temporal Identity Verification Method for Person-Action Instance Search in Movies
AUTHORS: Jingyao Yang ; Chao Liang ; Yanrui Niu ; Baojin Huang ; Zhongyuan Wang
CATEGORY: cs.CV [cs.CV, cs.IR, cs.MM]
HIGHLIGHT: Specifically, in the spatial dimension, we propose an identity consistency verification scheme to optimize the direct fusion score of person INS and action INS.
75, TITLE: MOST-GAN: 3D Morphable StyleGAN for Disentangled Face Image Manipulation
AUTHORS: SAFA C. MEDIN et. al.
CATEGORY: cs.CV [cs.CV, cs.AI, cs.GR, cs.LG, I.2.10]
HIGHLIGHT: In contrast, we propose a framework that a priori models physical attributes of the face such as 3D shape, albedo, pose, and lighting explicitly, thus providing disentanglement by design.
76, TITLE: Leveraging SE(3) Equivariance for Self-Supervised Category-Level Object Pose Estimation
AUTHORS: XIAOLONG LI et. al.
CATEGORY: cs.CV [cs.CV]
HIGHLIGHT: To reduce the huge amount of pose annotations needed for category-level learning, we propose for the first time a self-supervised learning framework to estimate category-level 6D object pose from single 3D point clouds.During training, our method assumes no ground-truth pose annotations, no CAD models, and no multi-view supervision.
77, TITLE: FREGAN : An Application of Generative Adversarial Networks in Enhancing The Frame Rate of Videos
AUTHORS: Rishik Mishra ; Neeraj Gupta ; Nitya Shukla
CATEGORY: cs.CV [cs.CV, cs.LG, eess.IV, I.2.1]
HIGHLIGHT: In this paper, we investigated the GAN model and proposed FREGAN for the enhancement of frame rate in videos.
78, TITLE: OctField: Hierarchical Implicit Functions for 3D Modeling
AUTHORS: JIA-HENG TANG et. al.
CATEGORY: cs.GR [cs.GR, cs.CV]
HIGHLIGHT: In this work, we present a learnable hierarchical implicit representation for 3D surfaces, coded OctField, that allows high-precision encoding of intricate surfaces with low memory and computational budget.
79, TITLE: Single-Item Fashion Recommender: Towards Cross-Domain Recommendations
AUTHORS: Seyed Omid Mohammadi ; Hossein Bodaghi ; Ahmad Kalhor
CATEGORY: cs.IR [cs.IR, cs.CV, cs.LG]
HIGHLIGHT: Still, many challenges lie ahead, and this study tries to tackle some.
80, TITLE: Nested Multiple Instance Learning with Attention Mechanisms
AUTHORS: Saul Fuster ; Trygve Eftest�l ; Kjersti Engan
CATEGORY: cs.LG [cs.LG, cs.CV]
HIGHLIGHT: Experiments in classical image datasets show that our proposed model provides high accuracy performance as well as spotting relevant instances on image regions.
81, TITLE: FC2T2: The Fast Continuous Convolutional Taylor Transform with Applications in Vision and Graphics
AUTHORS: Henning Lange ; J. Nathan Kutz
CATEGORY: cs.LG [cs.LG, cs.AI, cs.CV]
HIGHLIGHT: In this paper, we revisit the Taylor series expansion from a modern Machine Learning perspective.
82, TITLE: Predicting Atlantic Multidecadal Variability
AUTHORS: Glenn Liu ; Peidong Wang ; Matthew Beveridge ; Young-Oh Kwon ; Iddo Drori
CATEGORY: cs.LG [cs.LG, cs.CV, physics.ao-ph]
HIGHLIGHT: We use data from the Community Earth System Model 1 Large Ensemble Project, a state-of-the-art climate model with 3,440 years of data.
83, TITLE: Get Fooled for The Right Reason: Improving Adversarial Robustness Through A Teacher-guided Curriculum Learning Approach
AUTHORS: Anindya Sarkar ; Anirban Sarkar ; Sowrya Gali ; Vineeth N Balasubramanian
CATEGORY: cs.LG [cs.LG, cs.CR, cs.CV]
HIGHLIGHT: We propose a non-iterative method that enforces the following ideas during training.
84, TITLE: Towards The Generalization of Contrastive Self-Supervised Learning
AUTHORS: Weiran Huang ; Mingyang Yi ; Xuyang Zhao
CATEGORY: cs.LG [cs.LG, cs.AI, cs.CV, stat.ML]
HIGHLIGHT: To this end, we present a theoretical explanation of how contrastive self-supervised pre-trained models generalize to downstream tasks.
85, TITLE: Mastering Atari Games with Limited Data
AUTHORS: Weirui Ye ; Shaohuai Liu ; Thanard Kurutach ; Pieter Abbeel ; Yang Gao
CATEGORY: cs.LG [cs.LG, cs.AI, cs.CV, cs.RO]
HIGHLIGHT: We propose a sample efficient model-based visual RL algorithm built on MuZero, which we name EfficientZero.
86, TITLE: PIE: Pseudo-Invertible Encoder
AUTHORS: Jan Jetze Beitler ; Ivan Sosnovik ; Arnold Smeulders
CATEGORY: cs.LG [cs.LG, cs.CV]
HIGHLIGHT: We introduce new class of likelihood-based autoencoders with pseudo bijective architecture, which we call Pseudo Invertible Encoders.
87, TITLE: Distantly Supervised Semantic Text Detection and Recognition for Broadcast Sports Videos Understanding
AUTHORS: Avijit Shah ; Topojoy Biswas ; Sathish Ramadoss ; Deven Santosh Shah
CATEGORY: cs.MM [cs.MM, cs.CV, cs.IR, cs.LG, H.5.1; I.4.8; I.2.10; I.2.7; I.4.9]
HIGHLIGHT: In this work we study extremely accurate semantic text detection and recognition in sports clocks, and challenges therein. Finally, we share our computational architecture pipeline to scale this system in industrial setting and proposed a robust dataset for the same to validate our results.
88, TITLE: Hierarchical Deep Residual Reasoning for Temporal Moment Localization
AUTHORS: Ziyang Ma ; Xianjing Han ; Xuemeng Song ; Yiran Cui ; Liqiang Nie
CATEGORY: cs.MM [cs.MM, cs.CL, cs.CV, cs.IR]
HIGHLIGHT: Toward this end, we propose a Hierarchical Deep Residual Reasoning (HDRR) model, which decomposes the video and sentence into multi-level representations with different semantics to achieve a finer-grained localization.
89, TITLE: Focal Attention Networks: Optimising Attention for Biomedical Image Segmentation
AUTHORS: Michael Yeung ; Leonardo Rundo ; Evis Sala ; Carola-Bibiane Sch�nlieb ; Guang Yang
CATEGORY: eess.IV [eess.IV, cs.AI, cs.CV, math.OC]
HIGHLIGHT: In this paper, we investigate the role of the Focal parameter in modulating attention, revealing a link between attention in loss functions and networks.
90, TITLE: Simulating Realistic MRI Variations to Improve Deep Learning Model and Visual Explanations Using GradCAM
AUTHORS: MUHAMMAD ILYAS PATEL et. al.
CATEGORY: eess.IV [eess.IV, cs.AI, cs.CV]
HIGHLIGHT: We use a modified HighRes3DNet model for solving brain MRI volumetric landmark detection problem.
91, TITLE: TorchXRayVision: A Library of Chest X-ray Datasets and Models
AUTHORS: JOSEPH PAUL COHEN et. al.
CATEGORY: eess.IV [eess.IV, cs.AI, cs.CV]
HIGHLIGHT: TorchXRayVision: A Library of Chest X-ray Datasets and Models
92, TITLE: IGCN: Image-to-graph Convolutional Network for 2D/3D Deformable Registration
AUTHORS: Megumi Nakao ; Mitsuhiro Nakamura ; Tetsuya Matsuda
CATEGORY: eess.IV [eess.IV, cs.CV, cs.LG]
HIGHLIGHT: We propose an image-to-graph convolutional network that achieves deformable registration of a 3D organ mesh for a single-viewpoint 2D projection image.
93, TITLE: Redundancy Reduction in Semantic Segmentation of 3D Brain Tumor MRIs
AUTHORS: Md Mahfuzur Rahman Siddiquee ; Andriy Myronenko
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: In this work, we maintained an encoder-decoder based segmentation network, but focused on a modification of network training process that minimizes redundancy under perturbations.
94, TITLE: Functional Neural Networks for Parametric Image Restoration Problems
AUTHORS: Fangzhou Luo ; Xiaolin Wu ; Yanhui Guo
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: In this work, we propose a novel system called functional neural network (FuncNet) to solve a parametric image restoration problem with a single model.
95, TITLE: M2MRF: Many-to-Many Reassembly of Features for Tiny Lesion Segmentation in Fundus Images
AUTHORS: Qing Liu ; Haotian Liu ; Yixiong Liang
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: In this paper, we propose a many-to-many reassembly of features (M2MRF).
96, TITLE: Fetal MRI By Robust Deep Generative Prior Reconstruction and Diffeomorphic Registration: Application to Gestational Age Prediction
AUTHORS: LUCILIO CORDERO-GRANDE et. al.
CATEGORY: eess.IV [eess.IV, cs.CV, physics.med-ph, 92C50, 94A08, I.2.1; J.3]
HIGHLIGHT: Thus, in this paper we propose a deep generative prior for robust volumetric reconstructions integrated with a diffeomorphic volume to slice registration method.
97, TITLE: Dual Attention Network for Heart Rate and Respiratory Rate Estimation
AUTHORS: Yuzhuo Ren ; Braeden Syrnyk ; Niranjan Avadhanam
CATEGORY: eess.IV [eess.IV, cs.CV, eess.SP]
HIGHLIGHT: We propose a convolutional neural network which leverages spatial attention and channel attention, which we call it dual attention network (DAN) to jointly estimate heart rate and respiratory rate with camera video as input.
98, TITLE: Cross-Modality Fusion Transformer for Multispectral Object Detection
AUTHORS: Fang Qingyun ; Han Dapeng ; Wang Zhaokui
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: To fully exploit the different modalities, we present a simple yet effective cross-modality feature fusion approach, named Cross-Modality Fusion Transformer (CFT) in this paper.
99, TITLE: Incorporating Boundary Uncertainty Into Loss Functions for Biomedical Image Segmentation
AUTHORS: Michael Yeung ; Guang Yang ; Evis Sala ; Carola-Bibiane Sch�nlieb ; Leonardo Rundo
CATEGORY: eess.IV [eess.IV, cs.AI, cs.CV, math.OC]
HIGHLIGHT: In this paper, we propose the Boundary Uncertainty, which uses morphological operations to restrict soft labelling to object boundaries, providing an appropriate representation of uncertainty in ground truth labels, and may be adapted to enable robust model training where systematic manual segmentation errors are present.
100, TITLE: Calibrating The Dice Loss to Handle Neural Network Overconfidence for Biomedical Image Segmentation
AUTHORS: MICHAEL YEUNG et. al.
CATEGORY: eess.IV [eess.IV, cs.AI, cs.CV, math.OC]
HIGHLIGHT: In this study, we identify poor calibration as an emerging challenge of deep learning based biomedical image segmentation.
101, TITLE: Correlation Between Image Quality Metrics of Magnetic Resonance Images and The Neural Network Segmentation Accuracy
AUTHORS: RAJARAJESWARI MUTHUSIVARAJAN et. al.
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: Correlation Between Image Quality Metrics of Magnetic Resonance Images and The Neural Network Segmentation Accuracy
102, TITLE: Self-Verification in Image Denoising
AUTHORS: HUANGXING LIN et. al.
CATEGORY: eess.IV [eess.IV, cs.CV]
HIGHLIGHT: We devise a new regularization, called self-verification, for image denoising.
103, TITLE: Influential Prototypical Networks for Few Shot Learning: A Dermatological Case Study
AUTHORS: Ranjana Roy Chowdhury ; Deepti R. Bathula
CATEGORY: eess.IV [eess.IV, cs.CV, cs.LG]
HIGHLIGHT: In this work, we propose a novel version of PN that attributes weights to support samples corresponding to their influence on the support sample distribution.
104, TITLE: Unpaired Learning for High Dynamic Range Image Tone Mapping
AUTHORS: Yael Vinker ; Inbar Huberman-Spiegelglas ; Raanan Fattal
CATEGORY: eess.IV [eess.IV, cs.CV, cs.LG]
HIGHLIGHT: In this paper we describe a new tone-mapping approach guided by the distinct goal of producing low dynamic range (LDR) renditions that best reproduce the visual characteristics of native LDR images.
105, TITLE: Learning to Detect Open Carry and Concealed Object with 77GHz Radar
AUTHORS: XIANGYU GAO et. al.
CATEGORY: eess.SP [eess.SP, cs.CV]
HIGHLIGHT: In this paper, we focus on the relatively unexplored area of using low-cost 77GHz mmWave radar for the carried objects detection problem.