def visualize_mask_samples(dataset_train):
image_ids = [0, 1, 2, 3]
for image_id in image_ids:
image = dataset_train.load_image(image_id)
mask, class_ids = dataset_train.load_mask(image_id)
visualize.display_top_masks(image, mask, class_ids,
dataset_train.class_names)
def visualize_dataset(self, dataset, nbImages):
try:
image_ids = np.random.choice(dataset.image_ids, nbImages)
for image_id in image_ids:
image = dataset.load_image(image_id, 0)[:, :, :3]
mask, class_ids = dataset.load_mask(image_id)
visualize.display_top_masks(image, mask, class_ids,
dataset.class_names)
except Exception as e:
print('Error-Could not visualize dataset: ' + str(e))
def visualize_data(data):
image_ids = data.image_ids[:10]
for image_id in image_ids:
image = data.load_image(image_id)
mask, class_ids = data.load_mask(image_id)
#print(class_ids)
make_save("debug/%s" % image_id)
visualize.display_top_masks(image,
mask,
class_ids,
data.class_names,
limit=1)
dataset_train.load_shapes(500, config.IMAGE_SHAPE[0], config.IMAGE_SHAPE[1])
dataset_train.prepare()
# Validation dataset
dataset_val = ShapesDataset()
dataset_val.load_shapes(50, config.IMAGE_SHAPE[0], config.IMAGE_SHAPE[1])
dataset_val.prepare()
# In[6]:
# Load and display random samples
image_ids = np.random.choice(dataset_train.image_ids, 4)
for image_id in image_ids:
image = dataset_train.load_image(image_id)
mask, class_ids = dataset_train.load_mask(image_id)
visualize.display_top_masks(image, mask, class_ids,
dataset_train.class_names)
# ## Ceate Model
# In[ ]:
# Create model in training mode
model = modellib.MaskRCNN(mode="training", config=config, model_dir=MODEL_DIR)
# In[7]:
# Which weights to start with?
init_with = "coco" # imagenet, coco, or last
if init_with == "imagenet":
model.load_weights(model.get_imagenet_weights(), by_name=True)
dataset_train = ShapesDataset()
dataset_train.load_shapes(500, config.IMAGE_SHAPE[0], config.IMAGE_SHAPE[1])
dataset_train.prepare()
# Validation dataset
dataset_val = ShapesDataset()
dataset_val.load_shapes(50, config.IMAGE_SHAPE[0], config.IMAGE_SHAPE[1])
dataset_val.prepare()
# Load and display random samples
image_ids = np.random.choice(dataset_train.image_ids, 4)
for image_id in image_ids:
image = dataset_train.load_image(image_id)
mask, class_ids = dataset_train.load_mask(image_id)
visualize.display_top_masks(image, mask, class_ids, dataset_train.class_names, target=VIS_DIR+"rnd_smpl_" + str(image_id) + ".jpg" )
##########################################
###### CREATE MODEL ######
##########################################
# Create model in training mode
model = modellib.MaskRCNN(mode="training", config=config,
model_dir=MODEL_DIR)
# Which weights to start with?
init_with = "coco" # imagenet, coco, or last
if init_with == "imagenet":
dataset_train = NucleiDataset()
dataset_train.load_nuclei()
dataset_train.prepare()
dataset_val = NucleiDataset()
dataset_val.load_nuclei()
dataset_val.prepare()
# Load and display a test
for i in range(0,10):
TRAIN_PATH = '../stage1_train/'
image = dataset_train.load_image(i)
mask, class_ids = dataset_train.load_mask(i)
print(class_ids)
visualize.display_top_masks(image, mask, class_ids,['background','nuclei'])
# Create model in training mode
model = modellib.MaskRCNN(mode="training", config=config,
model_dir=MODEL_DIR)
# Which weights to start with?
init_with = "coco" # imagenet, coco, or last
if init_with == "imagenet":
model.load_weights(model.get_imagenet_weights(), by_name=True)
elif init_with == "coco":
# Load weights trained on MS COCO, but skip layers that
# are different due to the different number of classes
# See README for instructions to download the COCO weights
model.load_weights(COCO_MODEL_PATH, by_name=True,
data = MujocoData(None)
data.load(name)
data.load_mujoco(-1)
data.prepare()
return data
#just makes a basic mujoco dataset
if __name__ == '__main__':
fn = 'dataset.pickle'
if not os.path.exists(fn):
data = make_dataset(count=256)
data.save(fn)
else:
data = load_dataset(fn)
print('loaded dataset')
image_ids = data.image_ids[:10]
for image_id in image_ids:
image = data.load_image(image_id)
mask, class_ids = data.load_mask(image_id)
make_save("%s" % image_id)
visualize.display_top_masks(image,
mask,
class_ids,
data.class_names,
limit=2)
Validate,
learning_rate=config.LEARNING_RATE / 10,
epochs=100,
layers="all")
if 0:
image_ids = np.random.choice([
tid for tid in Train.image_ids if Train.load_mask(tid)[1][0] == 2
], 4)
for image_id in image_ids:
image = Train.load_image(image_id)
mask, class_ids = Train.load_mask(image_id)
print(mask.shape, class_ids, image.shape)
visualize.display_top_masks(image,
mask,
class_ids,
Train.class_names,
limit=len(class_ids))
class InferenceConfig(RSNAConfig):
GPU_COUNT = 1
IMAGES_PER_GPU = 1
inference_config = InferenceConfig()
if 0:
# Load trained weights
# Recreate the model in inference mode
model = modellib.MaskRCNN(mode="inference",
config=inference_config,
model_dir="./model")
