def __data_generation(self, img_ids_temp):
"Generates data containing batch_size samples"
# Initialization
X = np.empty((self.batch_size, *self.dim, self.n_channels))
y = np.empty((self.batch_size), dtype=int)
# Generate data
for i, ID in enumerate(img_ids_temp):
image = load_dicom_image(ID,
to_RGB=self.to_RGB,
rescale=self.rescale)
try:
image = Image.fromarray(image)
except:
print(
"Pil.Image can't read image. Possible 12 or 16 bit image. Try rescale=True to scale to 8 bit."
)
image = image.resize((self.dim[0], self.dim[1]))
X[i, ] = image
ann = self.imgs_anns_dict[ID][0]
y[i] = self.dataset.classes_dict[ann["labelId"]]["class_id"]
return X, to_categorical(y, num_classes=self.n_classes)
def test_visualize(p):
labels_dict = {
"L_egJRyg": 1, # bounding box
"L_MgevP2": 2, # polygon
"L_D21YL2": 3, # freeform
"L_lg7klg": 4, # line
"L_eg69RZ": 5, # location
"L_GQoaJg": 6, # global_image
"L_JQVWjZ": 7, # global_series
"L_3QEOpg": 8, # global_exam
}
p.set_labels_dict(labels_dict)
ct_dataset = p.get_dataset_by_id("D_qGQdpN")
ct_dataset.prepare()
# image with multiple annotations
image_id = ct_dataset.get_image_ids()[7]
grey_image = visualize.load_dicom_image(image_id)
rgb_image = visualize.load_dicom_image(image_id, to_RGB=True)
scaled_image_1 = visualize.load_dicom_image(image_id, rescale=True)
assert np.amax(grey_image) == 701
assert np.amax(scaled_image_1) == 255
assert grey_image.shape == (256, 256)
assert rgb_image.shape == (256, 256, 3)
scaled_image_2, gt_class_id, gt_bbox, gt_mask = visualize.get_image_ground_truth(
image_id, ct_dataset
)
assert len(gt_class_id) == len(gt_class_id)
assert gt_mask.shape == (256, 256, 5)
def __data_generation(self, img_ids_temp):
"Generates data containing batch_size samples"
# Initialization
X = np.empty((self.batch_size, *self.dim, self.n_channels))
y = np.empty((self.batch_size), dtype=int)
# Generate data
for i, ID in enumerate(img_ids_temp):
image = load_dicom_image(ID, to_RGB=True)
image = Image.fromarray(image)
image = image.resize((self.dim[0], self.dim[1]))
# image = resize(image, (self.dim[0], self.dim[1]))
X[i, ] = image
ann = self.imgs_anns_dict[ID][0]
y[i] = self.dataset.classes_dict[ann["labelId"]]["class_id"]
return X, to_categorical(y, num_classes=self.n_classes)
def create_tf_bbox_example(annotations, image_id, classes_dict):
image = visualize.load_dicom_image(image_id)
width = int(image.shape[1])
height = int(image.shape[0])
# TODO: I think object detection API decoder needs jpeg images...
raw_img = visualize.load_dicom_image(image_id, to_RGB=True)
img = Image.fromarray(raw_img)
img_buffer = io.BytesIO()
img.save(img_buffer, format="jpeg")
encoded_jpg = Image.open(img_buffer)
if encoded_jpg.format != "JPEG":
raise ValueError("Image format not JPEG")
key = hashlib.sha256(img_buffer.getvalue()).hexdigest()
xmins = [
] # List of normalized left x coordinates in bounding box (1 per box)
xmaxs = [
] # List of normalized right x coordinates in bounding box (1 per box)
ymins = [
] # List of normalized top y coordinates in bounding box (1 per box)
ymaxs = [
] # List of normalized bottom y coordinates in bounding box (1 per box)
classes_text = [] # List of string class name of bounding box (1 per box)
classes = [] # List of integer class id of bounding box (1 per box)
# per annotation
for a in annotations:
w = int(a["data"]["width"])
h = int(a["data"]["height"])
x_min = int(a["data"]["x"])
y_min = int(a["data"]["y"])
x_max = x_min + w
y_max = y_min + h
# WARN: these are normalized
xmins.append(float(x_min / width))
xmaxs.append(float(x_max / width))
ymins.append(float(y_min / height))
ymaxs.append(float(y_max / height))
classes_text.append(a["labelId"].encode("utf8"))
classes.append(classes_dict[a["labelId"]]["class_id"])
# print(classes)
tf_example = tf.train.Example(features=tf.train.Features(
feature={
"image/height":
dataset_util.int64_feature(height),
"image/width":
dataset_util.int64_feature(width),
"image/filename":
dataset_util.bytes_feature(image_id.encode("utf8")),
"image/source_id":
dataset_util.bytes_feature(image_id.encode("utf8")),
"image/key/sha256":
dataset_util.bytes_feature(key.encode("utf8")),
"image/encoded":
dataset_util.bytes_feature(img_buffer.getvalue()),
"image/format":
dataset_util.bytes_feature("jpg".encode("utf8")),
"image/object/bbox/xmin":
dataset_util.float_list_feature(xmins),
"image/object/bbox/xmax":
dataset_util.float_list_feature(xmaxs),
"image/object/bbox/ymin":
dataset_util.float_list_feature(ymins),
"image/object/bbox/ymax":
dataset_util.float_list_feature(ymaxs),
"image/object/class/text":
dataset_util.bytes_list_feature(classes_text),
"image/object/class/label":
dataset_util.int64_list_feature(classes),
}))
return tf_example
