I'm confused about the dim9,did the paper give a input dimension as 9?

[LeopoldACC]

the code show as below

def forward(self, x):
        batch_size = x.size(0)
        num_points = x.size(2)

        x = get_graph_feature(x, k=self.k, dim9=True)   # (batch_size, 9, num_points) -> (batch_size, 9*2, num_points, k)

[antao97]

DGCNN follows the setting of PointNet for semantic segmentation. In the 5.1. Applications of the PointNet paper, the authors introduce the 9-dim vector for semantic segmentation.

The description is presented as follows:

To prepare training data, we firstly split points by room, and then sample rooms into blocks with area 1m by 1m. We train our segmentation version of PointNet to predict per point class in each block. Each point is represented by a 9-dim vector of XYZ, RGB and normalized location as to the room (from 0 to 1).

[LeopoldACC]

thanks a lot!