one-hot function
Opened this issue · 3 comments
xuyingxiao commented
Hi,
Thanks for your re-implementation. It seems work well.
But I can't understand the one-hot function, it seems like a bitwise_not function, in line 9-10:
one_hot_label = torch.eye(n_classes, device=device, requires_grad=requires_grad)[label]
For binary segmentaion, what's the mean of tensor([[1.]], device='cuda:0', requires_grad=True)[label](where label is a 512*512 binary mask)
ashnair1 commented
A further question regarding one hot label I had was, why use
one_hot_label = torch.eye(
n_classes, device=device, requires_grad=requires_grad)[label]
when you could simply use
one_hot_label = F.one_hot(label, n_classes) ?
SonwYang commented
class BoundaryLoss(nn.Module):
"""Boundary Loss proposed in:
Alexey Bokhovkin et al., Boundary Loss for Remote Sensing Imagery Semantic Segmentation
https://arxiv.org/abs/1905.07852
"""
def __init__(self, theta0=3, theta=5):
super().__init__()
self.theta0 = theta0
self.theta = theta
def crop(self, w, h, target):
nt, ht, wt = target.size()
offset_w, offset_h = (wt - w) // 2, (ht - h) // 2
if offset_w > 0 and offset_h > 0:
target = target[:, offset_h:-offset_h, offset_w:-offset_w]
return target
def to_one_hot(self, target, size):
n, c, h, w = size
ymask = torch.FloatTensor(size).zero_()
new_target = torch.LongTensor(n, 1, h, w)
if target.is_cuda:
ymask = ymask.cuda(target.get_device())
new_target = new_target.cuda(target.get_device())
new_target[:, 0, :, :] = torch.clamp(target.detach(), 0, c - 1)
ymask.scatter_(1, new_target, 1.0)
return torch.autograd.Variable(ymask)
def forward(self, pred, gt):
"""
Input:
- pred: the output from model (before softmax)
shape (N, C, H, W)
- gt: ground truth map
shape (N, H, w)
Return:
- boundary loss, averaged over mini-bathc
"""
gt = torch.squeeze(gt)
n, c, h, w = pred.shape
log_p = F.log_softmax(pred, dim=1)
# softmax so that predicted map can be distributed in [0, 1]
pred = torch.softmax(pred, dim=1)
# one-hot vector of ground truth
gt = self.crop(w, h, gt)
one_hot_gt = self.to_one_hot(gt, log_p.size())
# boundary map
gt_b = F.max_pool2d(
1 - one_hot_gt, kernel_size=self.theta0, stride=1, padding=(self.theta0 - 1) // 2)
gt_b -= 1 - one_hot_gt
pred_b = F.max_pool2d(
1 - pred, kernel_size=self.theta0, stride=1, padding=(self.theta0 - 1) // 2)
pred_b -= 1 - pred
# extended boundary map
gt_b_ext = F.max_pool2d(
gt_b, kernel_size=self.theta, stride=1, padding=(self.theta - 1) // 2)
pred_b_ext = F.max_pool2d(
pred_b, kernel_size=self.theta, stride=1, padding=(self.theta - 1) // 2)
# reshape
gt_b = gt_b.view(n, c, -1)
pred_b = pred_b.view(n, c, -1)
gt_b_ext = gt_b_ext.view(n, c, -1)
pred_b_ext = pred_b_ext.view(n, c, -1)
# Precision, Recall
P = torch.sum(pred_b * gt_b_ext, dim=2) / (torch.sum(pred_b, dim=2) + 1e-7)
R = torch.sum(pred_b_ext * gt_b, dim=2) / (torch.sum(gt_b, dim=2) + 1e-7)
# Boundary F1 Score
BF1 = 2 * P * R / (P + R + 1e-7)
# summing BF1 Score for each class and average over mini-batch
loss = torch.mean(1 - BF1)
return loss
justalonelc commented
类边界损失（nn.模块）：“”“边界损失提出：Alexey Bokhovkin等人，遥感影像语义分割的边界损失 https://arxiv.org/abs/1905.07852
"""def __init__(self, theta0=3, theta=5): super().__init__() self.theta0 = theta0 self.theta = theta def crop(self, w, h, target): nt, ht, wt = target.size() offset_w, offset_h = (wt - w) // 2, (ht - h) // 2 if offset_w > 0 and offset_h > 0: target = target[:, offset_h:-offset_h, offset_w:-offset_w] return target def to_one_hot(self, target, size): n, c, h, w = size ymask = torch.FloatTensor(size).zero_() new_target = torch.LongTensor(n, 1, h, w) if target.is_cuda: ymask = ymask.cuda(target.get_device()) new_target = new_target.cuda(target.get_device()) new_target[:, 0, :, :] = torch.clamp(target.detach(), 0, c - 1) ymask.scatter_(1, new_target, 1.0) return torch.autograd.Variable(ymask) def forward(self, pred, gt): """ Input: - pred: the output from model (before softmax) shape (N, C, H, W) - gt: ground truth map shape (N, H, w) Return: - boundary loss, averaged over mini-bathc """ gt = torch.squeeze(gt) n, c, h, w = pred.shape log_p = F.log_softmax(pred, dim=1) # softmax so that predicted map can be distributed in [0, 1] pred = torch.softmax(pred, dim=1) # one-hot vector of ground truth gt = self.crop(w, h, gt) one_hot_gt = self.to_one_hot(gt, log_p.size()) # boundary map gt_b = F.max_pool2d( 1 - one_hot_gt, kernel_size=self.theta0, stride=1, padding=(self.theta0 - 1) // 2) gt_b -= 1 - one_hot_gt pred_b = F.max_pool2d( 1 - pred, kernel_size=self.theta0, stride=1, padding=(self.theta0 - 1) // 2) pred_b -= 1 - pred # extended boundary map gt_b_ext = F.max_pool2d( gt_b, kernel_size=self.theta, stride=1, padding=(self.theta - 1) // 2) pred_b_ext = F.max_pool2d( pred_b, kernel_size=self.theta, stride=1, padding=(self.theta - 1) // 2) # reshape gt_b = gt_b.view(n, c, -1) pred_b = pred_b.view(n, c, -1) gt_b_ext = gt_b_ext.view(n, c, -1) pred_b_ext = pred_b_ext.view(n, c, -1) # Precision, Recall P = torch.sum(pred_b * gt_b_ext, dim=2) / (torch.sum(pred_b, dim=2) + 1e-7) R = torch.sum(pred_b_ext * gt_b, dim=2) / (torch.sum(gt_b, dim=2) + 1e-7) # Boundary F1 Score BF1 = 2 * P * R / (P + R + 1e-7) # summing BF1 Score for each class and average over mini-batch loss = torch.mean(1 - BF1) return loss
thank you for your code