tvars, grads, predE, predA, predB, predK, l2loss, grad_predE, \
grad_predA, grad_predB, grad_predK, grad_l2loss, x, y_ = CNN_B(im_size, out_size, L,
batch_size=1,wd=0.01,layers=len(numfilt),numfilt=numfilt)
#Initialize CNN
optimizer = tf.train.AdamOptimizer(1e-5, epsilon=1e-7)
apply_gradients = optimizer.apply_gradients(zip(grads, tvars))
###########################################################################################
# DEFINE SNAKE INFERENCE
###########################################################################################
niter = 50
print('Creating snake inference graph...', flush=True)
with tf.device('/cpu:0'):
tf_u, tf_v, tf_du, tf_dv, tf_Du, tf_Dv, tf_u0, tf_v0, tf_du0, tf_dv0, \
tf_alpha, tf_beta, tf_kappa = snake_graph(out_size, L,niter=niter)
###########################################################################################
#Prepare folder to save network
###########################################################################################
print('Preparing model folder...', flush=True)
start_epoch = 0
if not os.path.isdir(model_path):
os.makedirs(model_path)
# Add ops to save and restore all the variables.
saver = tf.train.Saver()
###########################################################################################
# DEFINE EPOCH
]
epoch = -1
for s in modelnames:
epoch = max(int(s.split('-')[-1].split('.')[0]), epoch)
start_epoch = epoch + 1
# Add ops to save and restore all the variables.
saver = tf.train.Saver()
#Initialize CNN
optimizer = tf.train.AdamOptimizer(1e-4, epsilon=1e-6)
apply_gradients = optimizer.apply_gradients(zip(grads, tvars))
with tf.device('/cpu:0'):
tf_u, tf_v, tf_du, tf_dv, tf_Du, tf_Dv, tf_u0, tf_v0, tf_du0, tf_dv0, \
tf_alpha, tf_beta, tf_kappa = snake_graph(out_size, L)
def epoch(n, i, mode):
# mode (str): train or test
batch_ind = np.arange(i, i + batch_size)
batch = np.copy(images[:, :, :, batch_ind])
batch_mask = np.copy(masks[:, :, :, batch_ind])
thisGT = np.copy(GT[:, :, batch_ind[0]])
if mode is 'train':
ang = np.random.rand() * 360
for j in range(len(batch_ind)):
for b in range(batch.shape[2]):
batch[:, :, b, j] = imrotate(batch[:, :, b, j], ang)
batch_mask[:, :, 0, j] = imrotate(batch_mask[:, :, 0, j],
ang,
