CVPR2018_RotationNet: Joint Object Categorization and Pose Estimation Using Multiviews from Unsupervised Viewpoints

本文转载自查看原文 2018-03-27 16:44 1029 论文阅读笔记

Author

一作: Asako Kanezaki 是个漂亮的小姐姐， AIST 日本东京的研究所

Abstract

提出了RotationNet,用multi-view图片作为输入，同时估计物体的姿态和类别

它是用部分的multi-view图片做inference,这个性质使得它很实用,因为很多场景都只有部分的view

同时，论文提出的pose alignment的方法可以获得不同类别共享的的view-specific feature representations

RotationNet在ModelNet datasets 10- 40-class中都获得了SOTA

同时在object pose estimation中也获得了SOTA, 即使它在没有已知pose标签的数据集上训练

Introduction

However, in real-world scenarios, objects can often only be observed from limited viewpoints due to occlusions，which makes it difficult to rely on multi-view representations that are earned with the whole circumference（本文主要解决的问题: 当没有全围表示时，怎么用有限的view point images来学习特征）

object classification 和 viewpoint estimation时一个紧密相关的问题对，所以可以进行互补

不是很看的懂，挖坑待填(等CVPR accepted list出来，看大神解读)

免责声明！

本站转载的文章为个人学习借鉴使用，本站对版权不负任何法律责任。如果侵犯了您的隐私权益，请联系本站邮箱yoyou2525@163.com删除。

猜您在找 [CVPR 2019]Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation 论文阅读（Zhe Cao——【CVPR2017】Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields ） 6D姿态估计从0单排——看论文的小鸡篇——Learning 6D Object Pose Estimation using 3D Object Coordinates 【论文笔记】Learning to Estimate 3D Human Pose and Shape from a Single Color Image(CVPR 2018) CVPR2018_Frustum PointNets for 3D Object Detection from RGB-D Data 2019CVPR论文 HIgh Resolution Representation Learning for Human Pose Estimation代码解读【CVPR 2019】论文阅读：3D human pose estimation in video with temporal convolutions and semi-supervised training 【CVPR 2022】论文阅读：MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimation 论文阅读笔记七：Structure Inference Network:Object Detection Using Scene-Level Context and Instance-Level Relationships(CVPR2018) Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields（理解）