ZhouYuxuanYX / Hyperformer

This is the official implementation of our paper "Hypergraph Transformer for Skeleton-based Action Recognition."
Apache License 2.0
77 stars 15 forks source link

收录期刊或会议 #7

Closed Brian417-cup closed 1 year ago

Brian417-cup commented 1 year ago

你好,想问一下这篇工作最后是被哪个期刊或会议接收的吗,看到你的代码仓库上写的是 preprint版的

ZhouYuxuanYX commented 1 year ago

您好,抱歉我刚看到您的邮件,这篇工作正好是投稿的ICCV2023,现在还在rebuttal阶段,如果被录取了第一时间告诉您!

On Tue, 30 May 2023 at 11:02, Brian417-cup @.***> wrote:

你好,想问一下这篇工作最后是被哪个期刊或会议接收的吗,看到你的代码仓库上写的是 preprint版的

— Reply to this email directly, view it on GitHub https://github.com/ZhouYuxuanYX/Hyperformer/issues/7, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKOZHLA2WV2O7OZXBQZFP73XIVPMVANCNFSM6AAAAAAYTMF7BE . You are receiving this because you are subscribed to this thread.Message ID: @.***>

ZhouYuxuanYX commented 1 year ago

你好,想问一下这篇工作最后是被哪个期刊或会议接收的吗,看到你的代码仓库上写的是 preprint版的

这次ICCV没中,主要原因是和paperwithcode的leaderboard上很多文章分数相比没有绝对优势,但是其实leaderboard上很多工作没有完全公正地采用标准的实验设置进行训练,不公平的实验设置包括但不限于:1)融合了六个模态 2)用其它模态替代了性能较差的motion模态 3)采用了PoseC3D提出的基于HRNET热力图提取的2D版本skeleton数据集 4)采用了更高的时间帧数。因此本文将会在下个版本给出各种设置下的结果进行公正比较