lshiwjx / 2s-AGCN

Two-Stream Adaptive Graph Convolutional Networks for Skeleton-Based Action Recognition in CVPR19
Other
665 stars 180 forks source link

which file is the code for the adjacency matrix A B C #9

Open durianno1 opened 5 years ago

lshiwjx commented 5 years ago

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A, PA and A1

durianno1 commented 5 years ago

oh,such a interesting name

zahidzqj commented 5 years ago

A 是三维(3xNxN),图上的Ak是NxN,是怎样转变的,V是怎么处理的?

durianno1 commented 5 years ago

A 是三维(3xNxN),图上的Ak是NxN,是怎样转变的,V是怎么处理的?

我看好像是比论文里多讨论了,出入边的情况,分别进行了信息聚合再合并

durianno1 commented 5 years ago

what is dimension V?thank u

durianno1 commented 5 years ago

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A, PA and A1

what is dimension V?thank u

1zgh commented 5 years ago

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A,PA和A1

V是

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A,PA和A1

V是多少?谢谢

V 是25个关节点相互之间关联的程度把?我是这样认为的

durianno1 commented 5 years ago

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A,PA和A1

V是

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A,PA和A1

V是多少?谢谢

V 是25个关节点相互之间关联的程度把?我是这样认为的

关联程度是论文中的哪个概念?矩阵B和C不是已经求好了吗?还有就是好像代码考虑了出入边的问题,把A变成三维的了

1zgh commented 5 years ago

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A,PA和A1

V是

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A,PA和A1

V是多少?谢谢

V 是25个关节点相互之间关联的程度把?我是这样认为的

关联程度是论文中的哪个概念?矩阵B和C不是已经求好了吗?还有就是好像代码考虑了出入边的问题,把A变成三维的了

A three-dimensional is the relationship of neighborhood partition of nodes.If you look at formula four in the paper, V, I mean the similarity between nodes.

Lyndsey-Xing commented 4 years ago

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A,PA和A1

V是

https://github.com/lshiwjx/2s-AGCN/blob/de1b7e098226682816e5578f31ebd0613477a671/model/agcn.py#L95-L102

A,PA和A1

V是多少?谢谢

V 是25个关节点相互之间关联的程度把?我是这样认为的

关联程度是论文中的哪个概念?矩阵B和C不是已经求好了吗?还有就是好像代码考虑了出入边的问题,把A变成三维的了

这个矩阵C(即代码中的A1),它的shape为(16,25,25),然后A、B为(3,25,25),请问这里是怎么变换的呢