概述
编辑 | | AI速递
计算机视觉 3月28日
【1】 USB: Universal-Scale Object Detection Benchmark
标题:USB:万能物体检测基准
链接:https://arxiv.org/abs/2103.14027
【2】 Multi-Target Domain Adaptation via Unsupervised Domain Classification for Weather Invariant Object Detection
标题:天气不变目标检测中基于无监督领域分类的多目标域自适应
链接:https://arxiv.org/abs/2103.13970
【3】 Unmanned Aerial Vehicle Visual Detection and Tracking using Deep Neural Networks: A Performance Benchmark
标题:基于深度神经网络的无人机视觉检测与跟踪:一种性能基准
链接:https://arxiv.org/abs/2103.13933
【4】 Robust and Accurate Object Detection via Adversarial Learning
标题:基于对抗性学习的鲁棒准确目标检测
链接:https://arxiv.org/abs/2103.13886
【5】 Universal Representation Learning from Multiple Domains for Few-shot Classification
标题:基于多域的通用表示学习在Few-Shot分类中的应用
链接:https://arxiv.org/abs/2103.13841
【6】 3D3L: Deep Learned 3D Keypoint Detection and Description for LiDARs
标题:3D3L:激光雷达的深度学习三维关键点检测与描述
链接:https://arxiv.org/abs/2103.13808
【7】 Spatial-spectral Hyperspectral Image Classification via Multiple Random Anchor Graphs Ensemble Learning
标题:基于多随机锚图集成学习的空间光谱高光谱图像分类
链接:https://arxiv.org/abs/2103.13710
【8】 Frame-rate Up-conversion Detection Based on Convolutional Neural Network for Learning Spatiotemporal Features
标题:基于卷积神经网络的空时特征学习帧率上变频检测
链接:https://arxiv.org/abs/2103.13674
【9】 Gaussian Guided IoU: A Better Metric for Balanced Learning on Object Detection
标题:高斯引导借条:一种更好的目标检测均衡学习度量
链接:https://arxiv.org/abs/2103.13613
【10】 MetaAlign: Coordinating Domain Alignment and Classification for Unsupervised Domain Adaptation
标题:MetaAlign:协调无监督领域自适应的领域对齐和分类
链接:https://arxiv.org/abs/2103.13575
【11】 Deep Learning with robustness to missing data: A novel approach to the detection of COVID-19
标题:对缺失数据具有鲁棒性的深度学习:一种新的冠状病毒检测方法
链接:https://arxiv.org/abs/2103.13833
【12】 Explainability Guided Multi-Site COVID-19 CT Classification
标题:可探测性指导下的多部位冠状病毒CT分类
链接:https://arxiv.org/abs/2103.13677
【13】 3D Reasoning for Unsupervised Anomaly Detection in Pediatric WbMRI
标题:儿童WbMRI中无监督异常检测的三维推理
链接:https://arxiv.org/abs/2103.13497
[分割/语义相关]:
【1】 Video Instance Segmentation with a Propose-Reduce Paradigm
标题:基于提议-简化范式的视频实例分割
链接:https://arxiv.org/abs/2103.13746
【2】 Evidential fully convolutional network for semantic segmentation
标题:用于语义分割的证据完全卷积网络
链接:https://arxiv.org/abs/2103.13544
【3】 Contextual Information Enhanced Convolutional Neural Networks for Retinal Vessel Segmentation in Color Fundus Images
标题:上下文信息增强的卷积神经网络在彩色眼底图像视网膜血管分割中的应用
链接:https://arxiv.org/abs/2103.13622
[人脸相关]:
【1】 Hierarchical Deep CNN Feature Set-Based Representation Learning for Robust Cross-Resolution Face Recognition
标题:基于分层深度CNN特征集的鲁棒交叉分辨率人脸识别表示学习
链接:https://arxiv.org/abs/2103.13851
【2】 JDSR-GAN: Constructing A Joint and Collaborative Learning Network for Masked Face Super-Resolution
标题:JDSR-GAN:构建蒙面人脸超分辨率联合协作学习网络
链接:https://arxiv.org/abs/2103.13676
[GAN/对抗式/生成式相关]:
【1】 ScanGAN360: A Generative Model of Realistic Scanpaths for 360
∘
^{circ}
∘ Images
链接:https://arxiv.org/abs/2103.13922
【2】 Generative-Adversarial-Networks-based Ghost Recognition
标题:基于产生式-对抗性网络的鬼魂识别
链接:https://arxiv.org/abs/2103.13858
【3】 RA-BNN: Constructing Robust & Accurate Binary Neural Network to Simultaneously Defend Adversarial Bit-Flip Attack and Improve Accuracy
标题:RA-BNN:构造鲁棒准确的二进制神经网络同时防御敌意Bit-Flip攻击并提高准确率
链接:https://arxiv.org/abs/2103.13813
【4】 AttrLostGAN: Attribute Controlled Image Synthesis from Reconfigurable Layout and Style
标题:AttrLostGAN:基于可重构布局和样式的属性控制图像合成
链接:https://arxiv.org/abs/2103.13722
【5】 MCTSteg: A Monte Carlo Tree Search-based Reinforcement Learning Framework for Universal Non-additive Steganography
标题:MCTSteg:一种基于蒙特卡罗树搜索的通用非加性隐写强化学习框架
链接:https://arxiv.org/abs/2103.13689
【6】 THAT: Two Head Adversarial Training for Improving Robustness at Scale
标题:那就是:为提高规模健壮性而进行的两个头部对抗性训练
链接:https://arxiv.org/abs/2103.13612
【7】 Deepfake Forensics via An Adversarial Game
标题:通过对抗性游戏进行的深伪取证(Deepfac Forensics)
链接:https://arxiv.org/abs/2103.13567
【8】 Matched sample selection with GANs for mitigating attribute confounding
标题:减少属性混淆的GANS匹配样本选择
链接:https://arxiv.org/abs/2103.13455
[图像/视频检索]:
【1】 More Photos are All You Need: Semi-Supervised Learning for Fine-Grained Sketch Based Image Retrieval
标题:更多的照片就是你所需要的:基于细粒度草图的图像检索的半监督学习
链接:https://arxiv.org/abs/2103.13990
[行为/时空/光流/姿态/运动]:
【1】 The ThreeDWorld Transport Challenge: A Visually Guided Task-and-Motion Planning Benchmark for Physically Realistic Embodied AI
标题:ThreeDWorld交通挑战:物理上逼真的人工智能的视觉导引任务和运动规划基准
链接:https://arxiv.org/abs/2103.14025
【2】 GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learning
标题:GyroFlow:陀螺仪引导的无监督光流学习
链接:https://arxiv.org/abs/2103.13725
【3】 MBA-VO: Motion Blur Aware Visual Odometry
标题:MBA-VO:运动模糊视觉里程计
链接:https://arxiv.org/abs/2103.13684
【4】 STA-VPR: Spatio-temporal Alignment for Visual Place Recognition
标题:STA-VPR:视觉场所识别的时空对准
作者:Feng Lu,Baifan Chen,Xiang-Dong Zhou,Dezhen Song
链接:https://arxiv.org/abs/2103.13580
[半/弱/无监督相关]:
【1】 Self-Supervised Training Enhances Online Continual Learning
标题:自我指导培训促进在线持续学习
链接:https://arxiv.org/abs/2103.14010
【2】 Contrasting Contrastive Self-Supervised Representation Learning Models
标题:对比自监督表征学习模型
链接:https://arxiv.org/abs/2103.14005
【3】 Disentanglement-based Cross-Domain Feature Augmentation for Effective Unsupervised Domain Adaptive Person Re-identification
标题:基于解缠的跨域特征增强算法实现有效的无监督领域自适应人员再识别
链接:https://arxiv.org/abs/2103.13917
【4】 Inferring Latent Domains for Unsupervised Deep Domain Adaptation
标题:无监督深域自适应的潜在域推断
链接:https://arxiv.org/abs/2103.13873
【5】 Vectorization and Rasterization: Self-Supervised Learning for Sketch and Handwriting
标题:矢量化和光栅化:素描和笔迹的自我监督学习
链接:https://arxiv.org/abs/2103.13716
【6】 SSLayout360: Semi-Supervised Indoor Layout Estimation from 360
∘
^{circ}
∘ Panorama
链接:https://arxiv.org/abs/2103.13696
【7】 Contrast to Divide: Self-Supervised Pre-Training for Learning with Noisy Labels
标题:对比区分:带噪声标签学习的自我监督预训练(SSupervised Pre-Training for Learning With Noise Label)
链接:https://arxiv.org/abs/2103.13646
【8】 Rethinking Self-Supervised Learning: Small is Beautiful
标题:反思自主学习:小就是美
链接:https://arxiv.org/abs/2103.13559
【9】 DRANet: Disentangling Representation and Adaptation Networks for Unsupervised Cross-Domain Adaptation
标题:DRANet:无监督跨域自适应的解缠表示和自适应网络
链接:https://arxiv.org/abs/2103.13447
【10】 Semi-Supervised Learning for Bone Mineral Density Estimation in Hip X-ray Images
标题:半监督学习在髋部X线图像骨密度估计中的应用
链接:https://arxiv.org/abs/2103.13482
[跟踪相关]:
【1】 Tracking Pedestrian Heads in Dense Crowd
标题:在密集人群中跟踪行人头部
链接:https://arxiv.org/abs/2103.13516
【2】 A Framework for 3D Tracking of Frontal Dynamic Objects in Autonomous Cars
标题:一种自动驾驶汽车正面动态目标三维跟踪框架
链接:https://arxiv.org/abs/2103.13430
【3】 Zero-shot super-resolution with a physically-motivated downsampling kernel for endomicroscopy
标题:用于内窥镜检查的具有物理激励下采样核的零激发超分辨率
链接:https://arxiv.org/abs/2103.14015
[迁移学习/domain/主动学习/自适应]:
【1】 Boosting Binary Masks for Multi-Domain Learning through Affine Transformations
标题:利用仿射变换增强二值掩码的多域学习
作者:Massimiliano Mancini,Elisa Ricci,Barbara Caputo,Samuel Rota Buló
链接:https://arxiv.org/abs/2103.13894
【2】 OTCE: A Transferability Metric for Cross-Domain Cross-Task Representations
标题:OTCE:一种跨域跨任务表示的可移植性度量
链接:https://arxiv.org/abs/2103.13843
【3】 On Evolving Attention Towards Domain Adaptation
标题:关于对领域适应的关注的演变
链接:https://arxiv.org/abs/2103.13561
【4】 Addressing catastrophic forgetting for medical domain expansion
标题:解决医疗领域扩展的灾难性遗忘问题
链接:https://arxiv.org/abs/2103.13511
[裁剪/量化/加速相关]:
【1】 A Survey of Quantization Methods for Efficient Neural Network Inference
标题:高效神经网络推理的量化方法综述
链接:https://arxiv.org/abs/2103.13630
[超分辨率]:
【1】 Multi-frame Super-resolution from Noisy Data
标题:含噪数据的多帧超分辨率
链接:https://arxiv.org/abs/2103.13778
【2】 Asymmetric CNN for image super-resolution
标题:用于图像超分辨率的非对称CNN
链接:https://arxiv.org/abs/2103.13634
【3】 Designing a Practical Degradation Model for Deep Blind Image Super-Resolution
标题:一种实用的深盲图像超分辨率退化模型设计
链接:https://arxiv.org/abs/2103.14006
[其他视频相关]:
【1】 An Image is Worth 16x16 Words, What is a Video Worth?
标题:一张图片值16x16个字,视频值多少钱?
链接:https://arxiv.org/abs/2103.13915
【2】 Patch Craft: Video Denoising by Deep Modeling and Patch Matching
标题:补丁工艺:基于深度建模和补丁匹配的视频去噪
链接:https://arxiv.org/abs/2103.13767
【3】 TagMe: GPS-Assisted Automatic Object Annotation in Videos
标题:Tagme:GPS辅助的视频对象自动标注
链接:https://arxiv.org/abs/2103.13428
[其他]:
【1】 High-Fidelity Pluralistic Image Completion with Transformers
标题:用Transformer完成高保真的多元图像
链接:https://arxiv.org/abs/2103.14031
【2】 Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
标题:Swin Transformer:使用移位窗口的分层视觉转换器
链接:https://arxiv.org/abs/2103.14030
【3】 AutoLoss-Zero: Searching Loss Functions from Scratch for Generic Tasks
标题:AutoLoss-Zero:从零开始搜索一般任务的损失函数
链接:https://arxiv.org/abs/2103.14026
【4】 PlenOctrees for Real-time Rendering of Neural Radiance Fields
标题:神经辐射场实时绘制的PlenOctree算法
链接:https://arxiv.org/abs/2103.14024
【5】 AgentFormer: Agent-Aware Transformers for Socio-Temporal Multi-Agent Forecasting
标题:AgentFormer:面向社会时态多Agent预测的Agent感知转换器
链接:https://arxiv.org/abs/2103.14023
【6】 Orthogonal Projection Loss
标题:正交投影损耗
链接:https://arxiv.org/abs/2103.14021
【7】 Scaling-up Disentanglement for Image Translation
标题:用于图像平移的放大解缠
链接:https://arxiv.org/abs/2103.14017
【8】 Rethinking Deep Contrastive Learning with Embedding Memory
标题:嵌入记忆的深度对比学习再思考
链接:https://arxiv.org/abs/2103.14003
【9】 GridDehazeNet+: An Enhanced Multi-Scale Network with Intra-Task Knowledge Transfer for Single Image Dehazing
标题:GridDehazeNet+:一种支持任务内知识传递的增强型多尺度网络
链接:https://arxiv.org/abs/2103.13998
【10】 StyleLess layer: Improving robustness for real-world driving
标题:无样式层:提高真实驾驶的健壮性
链接:https://arxiv.org/abs/2103.13905
【11】 Progressive-X+: Clustering in the Consensus Space
标题:渐进式-X+:共识空间中的聚类
链接:https://arxiv.org/abs/2103.13875
【12】 Transform consistency for learning with noisy labels
标题:带噪声标签学习的变换一致性
链接:https://arxiv.org/abs/2103.13872
【13】 Group-CAM: Group Score-Weighted Visual Explanations for Deep Convolutional Networks
标题:GROUP-CAM:深卷积网络的组分数加权可视化解释
链接:https://arxiv.org/abs/2103.13859
【14】 I^3Net: Implicit Instance-Invariant Network for Adapting One-Stage Object Detectors
标题:I^3Net:适应一级目标检测器的隐式实例不变网络
链接:https://arxiv.org/abs/2103.13757
【15】 KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPs
链接:https://arxiv.org/abs/2103.13744
【16】 Spirit Distillation: Precise Real-time Prediction with Insufficient Data
标题:白酒蒸馏:在数据不足的情况下进行精确的实时预测
链接:https://arxiv.org/abs/2103.13733
【17】 ECINN: Efficient Counterfactuals from Invertible Neural Networks
标题:ECINN:来自可逆神经网络的有效反事实
链接:https://arxiv.org/abs/2103.13701
【18】 Learning Probabilistic Ordinal Embeddings for Uncertainty-Aware Regression
标题:不确定性感知回归的概率有序嵌入学习
链接:https://arxiv.org/abs/2103.13629
【19】 Exploiting Class Similarity for Machine Learning with Confidence Labels and Projective Loss Functions
标题:利用置信度标签和投影损失函数挖掘机器学习中的类相似性
链接:https://arxiv.org/abs/2103.13607
【20】 Recent Advances in Large Margin Learning
标题:大范围学习的最新进展
链接:https://arxiv.org/abs/2103.13598
【21】 Learning Dynamic Alignment via Meta-filter for Few-shot Learning
标题:基于元过滤的动态比对学习方法及其在少试性学习中的应用
链接:https://arxiv.org/abs/2103.13582
【22】 Test-Time Training for Deformable Multi-Scale Image Registration
标题:可变形多尺度图像配准的测试时间训练
链接:https://arxiv.org/abs/2103.13578
【23】 Efficient Feature Transformations for Discriminative and Generative Continual Learning
标题:区分产生式持续学习的有效特征变换
链接:https://arxiv.org/abs/2103.13558
【24】 Hierarchical Proxy-based Loss for Deep Metric Learning
标题:一种基于分层代理的损失深度度量学习方法
链接:https://arxiv.org/abs/2103.13538
【25】 A Broad Study on the Transferability of Visual Representations with Contrastive Learning
标题:基于对比学习的视觉表征迁移研究
链接:https://arxiv.org/abs/2103.13517
【26】 Projection: A Mechanism for Human-like Reasoning in Artificial Intelligence
标题:投影:人工智能中的一种类人推理机制
链接:https://arxiv.org/abs/2103.13512
【27】 A Survey of Multimedia Technologies and Robust Algorithms
标题:多媒体技术与鲁棒算法综述
链接:https://arxiv.org/abs/2103.13477
【28】 Diverse Branch Block: Building a Convolution as an Inception-like Unit
标题:多元化分支挡路:将卷积打造成一个类似盗梦空间的单位
链接:https://arxiv.org/abs/2103.13425
【29】 Foreground color prediction through inverse compositing
标题:通过逆合成进行前景颜色预测
链接:https://arxiv.org/abs/2103.13423
【30】 Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields
标题:MIP-NERF:一种抗混叠神经辐射场的多尺度表示
链接:https://arxiv.org/abs/2103.13415
【31】 Vision Transformers for Dense Prediction
标题:用于密度预测的视觉转换器
链接:https://arxiv.org/abs/2103.13413
【32】 Closing the Loop: Joint Rain Generation and Removal via Disentangled Image Translation
标题:闭合环路:通过解缠图像平移实现联合雨的产生和消除
链接:https://arxiv.org/abs/2103.13660
【33】 Artificial Intelligence in Tumor Subregion Analysis Based on Medical Imaging: A Review
标题:基于医学影像的人工智能在肿瘤亚区分析中的研究进展
链接:https://arxiv.org/abs/2103.13588
【34】 Task-Oriented Low-Dose CT Image Denoising
标题:面向任务的低剂量CT图像去噪
链接:https://arxiv.org/abs/2103.13557
最后
以上就是虚心哑铃为你收集整理的计算机视觉2021年3月28最新论文的全部内容,希望文章能够帮你解决计算机视觉2021年3月28最新论文所遇到的程序开发问题。
如果觉得靠谱客网站的内容还不错,欢迎将靠谱客网站推荐给程序员好友。
发表评论 取消回复