VPROMISE'S BLOG

SSH

SSH 的一些配置，方便远程连接，以及设置多设备跳转连接，完善VSCode远程连接服务器

Abbreviation of Journal and Conference

期刊和会议论文引用时缩写

learn shell

Shell 可以方便处理一些代码运行和进程，记录一下 Shell 的学习笔记

文献管理 Zotero

使用 Zotero 管理文献及相关配置，WebDAV同步，IOS客户端同步（待更新）

RSS to Telegram Bot

Flowerss bot for Telagram 通过 flowerss bot 搭建 Telegram RSS Bot, 最大的特点是支持应用内即时预览, 很方便

Temporal Convolutional Network (TCN)

《An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling》论文阅读笔记。这篇论文号称横扫序列模型中如RNN、GRU、LSTM等基本模型。

Transposed Convolution

本文翻译自《Up-sampling with Transposed Convolution》，作者FesianXu，这篇文章对转置卷积和反卷积有着很好的解释，这里将其翻译为中文

python */**

Python * 和 ** 语法, 转载于 https://blog.csdn.net/yilovexing/article/details/80577510

Regular expression

正则表达式学习, 转载于 https://github.com/cdoco/learn-regex-zh

Numpy-100

Numpy 100 exercises, 转载于 https://github.com/rougier/numpy-100

Dropout

dropout是CNN中防止过拟合提高效果的一个大杀器，但对于其为何有效，却众说纷纭，本文将基于两篇论文来讲解 dropout,转载于 https://blog.csdn.net/stdcoutzyx/article/details/49022443

3D HPM with SSL

This paper proposes a simple yet effective self-supervised correction mechanism to learn all intrinsic structures of human poses from abundant images

Cross View Fusion for 3D HPE

this paper present an approach to recover absolute 3D humanposes from multi-view images by incorporating multi-view geometric priors by USTC, MRA & TuSimple. code

替换NumPy数组中大于某个值的所有元素

问题:Python替换NumPy数组中大于某个值的所有元素，本文给出了几种思路

This work attempts to address the uncertainty of lifting the detected 2D joints to the 3D space by introduc-ing an intermediate state - Part-Centric Heatmap Triplets (HEMlets ), which shortens the gap between the 2D obser-vation and the 3D interpretation.

Deep learning based 2D HPE A survey

A Survey on Deep learning based 2D human pose estimation

Common Evaluation Metrics for HPE

Common Evaluation Metrics for Human Pose Estimation

My GitHub Starts

整理我 Star 的 GitHub 项目,便于查找，包括一些学习资料和教程，以及一些代码

GCP搭建v2ray

鉴于当前的网络环境,v2ray相对ss来说会稳定点,本文将介绍基于GCP搭建v2ray来实现科学上网

Related Parts Help HPE

Paper Reading: Does Learning Specific Features for Related Parts Help Human Pose Estimation? 在人体姿态估计任务中通过学习相关部位的特定特征来提高人体姿态估计的精度。

Multi-views 3D HPE

Paper Reading: Fast and Robust Multi-Person 3D Pose Estimation from Multiple Views. 本论文由浙大CAD实验室三维视觉研究组提出，完成了快速且鲁棒的多视角下多人三维姿态估计，该方法比之前的最先进的方法在相应的基准数据集上效果有非常明显的提升。

Deep HRNet for HPE

Paper Reading: Deep High-Resolution Representation Learning for Human Pose Estimation. 本文由中国科学技术大学和微软亚洲研究院视觉计算组完成，提出高分辨率深度神经网络（HRNet），与上篇论文 Multi-Person Pose Estimation with Enhanced Channel-wise and Spatial Information 并列为今年 Look Into Person (LIP) 竞赛 Single-Person Human Pose Estimation 挑战赛的冠军。