def upload_study_me(file_path, host, port):
file_dict = []
headers = {'Content-Type': 'multipart/related; '}
request_json = {
'request': 'post',
'route': '/',
'inference_command': 'get-bounding-box-2d'
}
images = load_image_data(file_path)
images = sort_images(images)
width = 0
height = 0
count = 0
for image in images:
try:
dcm_file = pydicom.dcmread(image.path)
if width == 0 or height == 0:
width = dcm_file.Columns
height = dcm_file.Rows
count += 1
field = str(count)
fo = open(image.path, 'rb').read()
filename = os.path.basename(os.path.normpath(image.path))
file_dict.append((field, (filename, fo, 'application/dicom')))
except:
print('File {} is not a DICOM file'.format(image.path))
continue
print('Sending {} files...'.format(count))
request_json['depth'] = count
request_json['height'] = height
request_json['width'] = width
file_dict.insert(
0,
('request_json',
('request', json.dumps(request_json).encode('utf-8'), 'text/json')))
me = MultipartEncoder(fields=file_dict)
boundary = me.content_type.split('boundary=')[1]
headers['Content-Type'] = headers['Content-Type'] + 'boundary="{}"'.format(
boundary)
r = requests.post('http://' + host + ':' + port + '/',
data=me,
headers=headers)
if r.status_code != 200:
print("Got error status code ", r.status_code)
exit(1)
multipart_data = decoder.MultipartDecoder.from_response(r)
json_response = json.loads(multipart_data.parts[0].text)
print("JSON response:", json_response)
def _get_images_and_masks(dicom_images, inference_results):
if isinstance(dicom_images, str):
images = load_image_data(dicom_images)
images = sort_images(images)
else:
images = dicom_images
if isinstance(inference_results, str):
masks = [np.fromfile(inference_results, dtype=np.uint8)]
else:
masks = inference_results
return (images, masks)
def upload_study_me(file_path,
model_type,
host,
port,
output_folder,
attachments,
override_inference_command=None,
send_study_size=False):
file_dict = []
headers = {'Content-Type': 'multipart/related; '}
images = load_image_data(file_path)
images = sort_images(images)
if model_type == BOUNDING_BOX:
print("Performing bounding box prediction")
inference_command = 'get-bounding-box-2d'
elif model_type == SEGMENTATION_MODEL:
if images[0].position is None:
# No spatial information available. Perform 2D segmentation
print("Performing 2D mask segmentation")
inference_command = 'get-probability-mask-2D'
else:
print("Performing 3D mask segmentation")
inference_command = 'get-probability-mask-3D'
else:
inference_command = 'other'
if override_inference_command:
inference_command = override_inference_command
request_json = {
'request': 'post',
'route': '/',
'inference_command': inference_command
}
count = 0
width = 0
height = 0
for att in attachments:
count += 1
field = str(count)
fo = open(att, 'rb').read()
filename = os.path.basename(os.path.normpath(att))
file_dict.append((field, (filename, fo, 'application/octet-stream')))
for image in images:
try:
dcm_file = pydicom.dcmread(image.path)
if width == 0 or height == 0:
width = dcm_file.Columns
height = dcm_file.Rows
count += 1
field = str(count)
fo = open(image.path, 'rb').read()
filename = os.path.basename(os.path.normpath(image.path))
file_dict.append((field, (filename, fo, 'application/dicom')))
except:
print('File {} is not a DICOM file'.format(image.path))
continue
print('Sending {} files...'.format(len(images)))
if send_study_size:
request_json['depth'] = count
request_json['height'] = height
request_json['width'] = width
file_dict.insert(
0,
('request_json',
('request', json.dumps(request_json).encode('utf-8'), 'text/json')))
me = MultipartEncoder(fields=file_dict)
boundary = me.content_type.split('boundary=')[1]
headers['Content-Type'] = headers['Content-Type'] + 'boundary="{}"'.format(
boundary)
r = requests.post('http://' + host + ':' + port + '/',
data=me,
headers=headers)
if r.status_code != 200:
print("Got error status code ", r.status_code)
exit(1)
multipart_data = decoder.MultipartDecoder.from_response(r)
json_response = json.loads(multipart_data.parts[0].text)
print("JSON response:", json_response)
if model_type == SEGMENTATION_MODEL:
mask_count = len(json_response["parts"])
# Assert that we get one binary part for each object in 'parts'
# The additional two multipart object are: JSON response and request:response digests
assert mask_count == len(multipart_data.parts) - 2, \
"The server must return one binary buffer for each object in `parts`. Got {} buffers and {} 'parts' objects" \
.format(len(multipart_data.parts) - 2, mask_count)
masks = [
np.frombuffer(p.content, dtype=np.uint8)
for p in multipart_data.parts[1:mask_count + 1]
]
if images[0].position is None:
# We must sort the images by their instance UID based on the order of the response:
identifiers = [
part['dicom_image']['SOPInstanceUID']
for part in json_response["parts"]
]
filtered_images = []
for id in identifiers:
image = next((img for img in images if img.instanceUID == id))
filtered_images.append(image)
test_inference_mask.generate_images_for_single_image_masks(
filtered_images, masks, json_response, output_folder)
else:
test_inference_mask.generate_images_with_masks(
images, masks, json_response, output_folder)
print("Segmentation mask images generated in folder: {}".format(
output_folder))
print("Saving output masks to files '{}/output_masks_*.npy".format(
output_folder))
for index, mask in enumerate(masks):
mask.tofile('{}/output_masks_{}.npy'.format(
output_folder, index + 1))
elif model_type == BOUNDING_BOX:
boxes = json_response['bounding_boxes_2d']
test_inference_boxes.generate_images_with_boxes(
images, boxes, output_folder)
with open(os.path.join(output_folder, 'response.json'), 'w') as outfile:
json.dump(json_response, outfile)
def upload_study_me(file_path, is_segmentation_model, host, port):
file_dict = []
headers = {'Content-Type': 'multipart/related; '}
images = load_image_data(file_path)
images = sort_images(images)
if not is_segmentation_model:
inference_command = 'get-bounding-box-2d'
elif images[0].position is None:
# No spatial information available. Perform 2D segmentation
inference_command = 'get-probability-mask-2D'
else:
inference_command = 'get-probability-mask-3D'
inference_command = "covid19"
request_json = {
'request': 'post',
'route': '/',
'inference_command': inference_command
}
width = 0
height = 0
count = 0
for image in images:
try:
dcm_file = pydicom.dcmread(image.path)
if width == 0 or height == 0:
width = dcm_file.Columns
height = dcm_file.Rows
count += 1
field = str(count)
fo = open(image.path, 'rb').read()
filename = os.path.basename(os.path.normpath(image.path))
file_dict.append((field, (filename, fo, 'application/dicom')))
except:
print('File {} is not a DICOM file'.format(image.path))
continue
print('Sending {} files...'.format(count))
request_json['depth'] = count
request_json['height'] = height
request_json['width'] = width
file_dict.insert(
0,
('request_json',
('request', json.dumps(request_json).encode('utf-8'), 'text/json')))
me = MultipartEncoder(fields=file_dict)
boundary = me.content_type.split('boundary=')[1]
headers['Content-Type'] = headers['Content-Type'] + 'boundary="{}"'.format(
boundary)
r = requests.post('http://' + host + ':' + port + '/',
data=me,
headers=headers)
if r.status_code != 200:
print("Got error status code ", r.status_code)
exit(1)
multipart_data = decoder.MultipartDecoder.from_response(r)
json_response = json.loads(multipart_data.parts[0].text)
print("JSON response:", json_response)
mask_count = len(json_response["parts"])
masks = [
np.frombuffer(p.content, dtype=np.uint8)
for p in multipart_data.parts[1:mask_count + 1]
]
if is_segmentation_model:
output_folder = 'output'
if images[0].position is None:
# We must sort the images by their instance UID based on the order of the response:
identifiers = [
part['dicom_image']['SOPInstanceUID']
for part in json_response["parts"]
]
filtered_images = []
for id in identifiers:
image = next((img for img in images if img.instanceUID == id))
filtered_images.append(image)
test_inference_mask.generate_images_for_single_image_masks(
filtered_images, masks, output_folder)
else:
test_inference_mask.generate_images_with_masks(
images, masks, output_folder)
print("Segmentation mask images generated in folder: {}".format(
output_folder))
print("Saving output masks to files 'output/output_masks_*.npy")
for index, mask in enumerate(masks):
mask.tofile('output/output_masks_{}.npy'.format(index + 1))
def worker(args, q_ind, q_value, r_ind, r_value):
fd = load_object_detector(args.object_detector)
dataset = "data/TinyTLP/"
all_directories = get_directories(dataset)
results = {}
for directory in all_directories:
try:
results_dir = results[directory] = []
plotted = results[f"{directory}_plots"] = []
iou = results[f"{directory}_iou"] = []
errors = results[f"{directory}_errors"] = []
ground_truth_file = dataset + directory + "/groundtruth_rect.txt"
images_wildcard = dataset + directory + "/img/*.jpg"
images_filelist = glob(images_wildcard)
images_filelist = sort_images(images_filelist)
de = None
if args.extract_density:
# de = DensityExtractor(images_filelist[0], args)
# de.create_grid()
de = GaussianDensityExtractor(images_filelist[0], args)
de.create_grid()
# Extract all ground truths
ground_truth, gt_measurements = extract_all_groud_truths(
ground_truth_file)
# Create PF
w, h = map(int, ground_truth[0][3:5])
pf = create_particle_filter(ground_truth[0], q_value, r_value,
images_filelist[0], args)
# Create images iterator
t = create_iterator(images_filelist, args)
# Create data association
da = MLPFAssociation(states=pf.S,
R=r_value,
H=pf.H,
threshold=args.outlier_detector_threshold)
# Iterate of every image
features = None
for i, im in enumerate(t):
img = read_img(im)
# Do prediction
pf.predict()
if i % args.resample_every == 0:
# Compute features
features = np.array(fd.compute_features(img))
if len(features) > 0:
# Do data association
if args.point_estimate:
features = [
np.array([i[0] + w / 2, i[1] + h / 2])
for i in features
]
psi, outlier, c = da.associate(features)
pf.update(psi, outlier, c)
else:
pf.assign_predicted()
else:
pf.assign_predicted()
gt = list(map(int, ground_truth[i]))
x = pf.get_x().T
# Plot features
if args.should_plot:
plot_result = plot(args, x, de, img, gt, features)
if args.show_plots is False:
plotted.append(plot_result)
if args.extract_density:
if args.should_plot is False:
de.estimate(x)
if args.point_estimate:
v = np.linalg.norm([
de.xtmean - gt[1] - gt[3] / 2,
de.ytmean - gt[2] - gt[4] / 2
],
axis=0)
results_dir.append(v)
else:
v = np.linalg.norm([
de.xtmean - gt[1], de.ytmean - gt[2],
(de.xbmean - de.xtmean) - gt[3],
(de.ybmean - de.ytmean) - gt[4]
])
t.set_description(f"Last error: {v:3.4}")
errors.append(v)
if not args.point_estimate:
iou.append(
get_iou(
[de.xtmean, de.ytmean, de.xbmean, de.ybmean],
[gt[1], gt[2], gt[1] + gt[3], gt[2] + gt[4]]))
results_dir.append([
de.xtmean, de.ytmean, de.xbmean - de.xtmean,
de.ybmean - de.ytmean
])
except Exception as e:
print(f"Crashed with error: {str(e)}. Q: {q_value}, R: {r_value}")
if args.point_estimate:
results_file = f"results/pf_R_{r_ind}_Q_{q_ind}_{args.object_detector}.pickle"
else:
results_file = f"results/pf_box_R_{r_ind}_Q_{q_ind}_{args.object_detector}.pickle"
with open(results_file, 'wb') as fp:
pickle.dump(results, fp)
def test_bkf_Q(dataset, all_directories, args):
results = {}
l = 2
R_value = 1
# Get images list from dataset
dataset = "data/TinyTLP/"
all_directories = get_directories(dataset)
fd = load_object_detector(args.object_detector)
for k, Q_value in enumerate(Q_values):
for dir in all_directories:
iou = []
results_dir = results[dir] = []
ground_truth_file = dataset + dir + "/groundtruth_rect.txt"
images_wildcard = dataset + dir + "/img/*.jpg"
images_filelist = glob(images_wildcard)
images_filelist = sort_images(images_filelist)
# Extract all ground truths
ground_truth, gt_measurements = extract_all_groud_truths(
ground_truth_file)
# Create KF
kf = create_box_kalman_filter(ground_truth[0], Q_value, R_value)
# Iterate of every image
t = tqdm(images_filelist[1:], desc="Processing")
da = DataAssociation(R=kf.R, H=kf.H, threshold=1)
# Create plot if should_plot is true
if args.should_plot:
fig, ax = create_fig_ax()
features = {}
for i, im in enumerate(t):
img = read_img(images_filelist[i])
# Compute features
features[i] = np.array(fd.compute_features(img))
# Do prediction
mu_bar, Sigma_bar = kf.predict()
# Do data association
da.update_prediction(mu_bar, Sigma_bar)
m = da.associate(features[i])
kf.update(m)
gt = list(map(int, ground_truth[i]))
kf_x = kf.get_x()
if args.should_plot:
plot_box_kf_result(ax, i, img, m, kf_x, ground_truth)
iou.append(
get_iou([kf_x[0], kf_x[1], kf_x[2], kf_x[3]],
[gt[1], gt[2], gt[1] + gt[3], gt[2] + gt[4]]))
results_dir.append(
[kf_x[0], kf_x[1], kf_x[2] - kf_x[0], kf_x[3] - kf_x[1]])
print(f"Dataset: {dir}, IoU: {np.mean(iou), np.std(iou)}")
with open(
f"results/kalman_filter_box_R_{l}_Q_{k}_{args.object_detector}.pickle",
'wb') as fp:
pickle.dump(results, fp)
