Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

LORE 这部分代码问题 #186

Open
cqray1990 opened this issue Oct 23, 2024 · 0 comments
Open

LORE 这部分代码问题 #186

cqray1990 opened this issue Oct 23, 2024 · 0 comments

Comments

@cqray1990
Copy link

wh labels计算的是:

   wh[k] = ct[0] - 1. * CorNer[0], ct[1] - 1. * CorNer[1], \
            ct[0] - 1. * CorNer[2], ct[1] - 1. * CorNer[3], \
            ct[0] - 1. * CorNer[4], ct[1] - 1. * CorNer[5], \
            ct[0] - 1. * CorNer[6], ct[1] - 1. * CorNer[7]

中心点减去四个角点的值,这个是宽高wh?

后面又通过 ind 求出中心点对应在特征图上的坐标xs,ys

再减去wh 得到是什么值?

def _get_wh_feat(ind, output, ttype):

width = output['hm'].shape[2]
xs = (ind % width).unsqueeze(2).int().float()
ys = (ind // width).unsqueeze(2).int().float()
if ttype == 'gt':
wh = output['wh']
elif ttype == 'pred':
wh = _tranpose_and_gather_feat(output['wh'], ind)
ct = torch.cat([xs, ys, xs, ys, xs, ys, xs, ys], dim=2)
bbx = ct - wh

return bbx

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant