def eval_net(net, dataset, gpu=False):
tot = 0
for i, b in enumerate(dataset):
X = b[0]
y = b[1]
X = torch.FloatTensor(X).unsqueeze(0)
y = torch.ByteTensor(y).unsqueeze(0)
if gpu:
X = Variable(X, volatile=True).cuda()
y = Variable(y, volatile=True).cuda()
else:
X = Variable(X, volatile=True)
y = Variable(y, volatile=True)
y_pred = net(X)
y_pred = (F.sigmoid(y_pred) > 0.6).float()
# y_pred = F.sigmoid(y_pred).float()
dice = dice_coeff(y_pred, y.float()).data[0]
tot += dice
print("i=", i)
return tot / i
def eval_net(net, dataset, gpu=False):
tot = 0
for i, b in enumerate(dataset):
X = b[0]
y = b[1]
X = torch.FloatTensor(X).unsqueeze(0)
y = torch.ByteTensor(y).unsqueeze(0)
if gpu:
X = Variable(X, volatile=True).cuda()
y = Variable(y, volatile=True).cuda()
else:
X = Variable(X, volatile=True)
y = Variable(y, volatile=True)
y_pred = net(X)
y_pred = (F.sigmoid(y_pred) > 0.6).float()
print('X shape: ', X.shape)
print('y_pred shape: ', y_pred.shape)
dice = dice_coeff(y_pred, y.float()).data[0]
tot += dice
if 0:
X = X.data.squeeze(0).cpu().numpy()
X = np.transpose(X, axes=[1, 2, 0])
y = y.data.squeeze(0).cpu().numpy()
y_pred = y_pred.data.squeeze(0).squeeze(0).cpu().numpy()
print(y_pred.shape)
fig = plt.figure()
ax1 = fig.add_subplot(1, 4, 1)
ax1.imshow(X)
ax2 = fig.add_subplot(1, 4, 2)
ax2.imshow(y)
ax3 = fig.add_subplot(1, 4, 3)
ax3.imshow((y_pred > 0.5))
Q = dense_crf(((X * 255).round()).astype(np.uint8), y_pred)
ax4 = fig.add_subplot(1, 4, 4)
print(Q)
ax4.imshow(Q > 0.5)
plt.show()
print("i=", i)
return tot / i
def eval_net(net, dataset, gpu=False):
tot = 0
mean_iou = 0
mean_precision = 0
mean_recall = 0
for i, b in enumerate(dataset):
X = b[0]
y = b[1]
X = torch.FloatTensor(X).unsqueeze(0)
y = torch.ByteTensor(y).unsqueeze(0)
if gpu:
X = Variable(X, volatile=True).cuda()
y = Variable(y, volatile=True).cuda()
else:
X = Variable(X, volatile=True)
y = Variable(y, volatile=True)
y_pred = net(X)
y_pred = (F.sigmoid(y_pred) > 0.6).float()
dice = dice_coeff(y_pred, y.float()).data[0]
tot += dice
# print('y_pred.data shape: ', y_pred.data.squeeze(0).squeeze(0).cpu().numpy().shape)
# print('y.data: ', y.data.squeeze(0).cpu().numpy().shape)
y = y.data.squeeze(0).cpu().numpy()
y = (y > 0.6).astype(np.float32)
y_pred = y_pred.data.squeeze(0).squeeze(0).cpu().numpy()
y_pred = (y_pred > 0.5).astype(np.float32)
# print('y.shape: ', y.shape)
# print('y_pred.shape: ', y_pred.shape)
# print('y binary : ', np.array_equal(y, y.astype(bool)))
# print('y_pred binary : ', np.array_equal(y_pred, y_pred.astype(bool)))
iou_val = get_iou(y_pred, y)
mean_iou += iou_val
precision_val = get_precision(y_pred, y)
mean_precision += precision_val
recall_val = get_recall(y_pred, y)
mean_recall += recall_val
if 0:
X = X.data.squeeze(0).cpu().numpy()
X = np.transpose(X, axes=[1, 2, 0])
# y = y.data.squeeze(0).cpu().numpy()
# y_pred = y_pred.data.squeeze(0).squeeze(0).cpu().numpy()
fig = plt.figure()
ax1 = fig.add_subplot(1, 4, 1)
ax1.imshow(X)
ax2 = fig.add_subplot(1, 4, 2)
ax2.imshow(y)
ax3 = fig.add_subplot(1, 4, 3)
ax3.imshow(y_pred)
Q = dense_crf(((X * 255).round()).astype(np.uint8), y_pred)
ax4 = fig.add_subplot(1, 4, 4)
print(Q)
ax4.imshow(Q > 0.5)
plt.show()
print("i=", i)
return tot / i, mean_iou / (i + 1.0), mean_precision / (
i + 1.0), mean_recall / (i + 1.0)