def test_main_process(self):
# pull test data from gitlab
input_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'b7abe78697075dc2e47de5327c9f2de3/cv_classifier_input_list.json'
input_list = lab_tools.pull_json_from_gitlab(input_url)
output_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'c9c8c98992abdf43036eed69d0e5fc3f/cv_output_list.json'
output_list = lab_tools.pull_json_from_gitlab(output_url)
# initialization
classifier_ = ClassifierSimpleCV()
classifier_.param["bkg_ref_path"] = '/home/shared/DT42/test_data/'\
'test_demo_app_with_calibration/ref_bkg.png'
image_list = []
for i in range(len(input_list)):
image_list.append(image.read_img(input_list[i]))
# run classifier
classifier_.reset()
classifier_.input_data = image_list
classifier_.run()
# compare results with reference
for j in range(len(classifier_.results)):
ref_data = output_list[j]['annotations'][0]['label']
tar_data = classifier_.results[j]['annotations'][0]['label']
self.assertEqual(ref_data, tar_data)
def __init__(self, dyda_config_path="", param=None):
"""Initialization function of dyda component.
This is called when an object is created from the class and
it allows the class to initialize the attributes of a class.
"""
super(ClassifierFrameDiff,
self).__init__(dyda_config_path=dyda_config_path)
class_name = self.__class__.__name__
self.set_param(class_name, param=param)
self.check_param_keys(["bkg_path", "thre"])
self.bkg_path = self.param["bkg_path"]
self.thre = self.param["thre"]
self.sel_min = 0
if "sel_min" in self.param.keys():
self.sel_min = self.param["sel_min"]
self.sel_max = -1
if "sel_max" in self.param.keys():
self.sel_max = self.param["sel_max"]
self.mean_axis = 1
if "mean_axis" in self.param.keys():
self.mean_axis = self.param["mean_axis"]
try:
self.bkg_img = image.read_img(self.bkg_path)
except:
self.logger.error("Fail to read background image %s" %
self.bkg_path)
sys.exit(0)
def test_main_process(self):
# pull test data from gitlab
config_url = ("https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/"
"dbd906035055b99da05308aa927b9f4a/DetectorYOLO.config")
dyda_config = lab_tools.pull_json_from_gitlab(config_url)
ref_url = ("https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/"
"6ce02b726d96edaf4cb4e4108fb9ba9f/DetectorYOLO_output.json")
ref_data = lab_tools.pull_json_from_gitlab(ref_url)
detector_yolo = DetectorYOLO(dyda_config_path=dyda_config)
input_path = '/home/shared/DT42/test_data/test_detector/dog.jpg'
detector_yolo.input_data = [image.read_img(input_path)]
detector_yolo.run()
tar_data = detector_yolo.results[0]
report = dict_comparator.get_diff(ref_data, tar_data)
self.assertEqual(report['extra_field'], [])
self.assertEqual(report['missing_field'], [])
# Since the filename depending on time stamp, it wont match anyway
for mis_match in report['mismatch_val']:
for tar_key in mis_match['tar_key']:
if tar_key == 'filename':
continue
else:
self.assertEqual(report['mismatch_val'], [])
def read_image(self, input_data):
try:
img_data = image.read_img(input_data, log=False)
if img_data is None or img_data.shape[0] == 0:
print('[dyda] Error: Cannot read %s' % input_data)
self.terminate_flag = True
return img_data
except Exception:
traceback.print_exc(file=sys.stdout)
self.terminate_flag = True
def set_config(self, input_data, config_data, use_key):
self.input_data = input_data
self.config_data = config_data
self.draw_tool.config_data = config_data
self.use_key = use_key
self.output_data = self.input_data
img_path = self.input_data['img_path']
self.init_img = image.read_img(img_path)
img_sz = self.init_img.shape
img_sz = np.multiply([img_sz[1], img_sz[0]], 1.0)
self.disp_scale = np.divide(img_sz, self.disp_sz)
def __init__(self, dyda_config_path=''):
""" __init__ of ClassifierSimpleCV
Variables:
input_data: a list of image array
results: defined by lab_tools.output_pred_classification (dict)
"""
super(ClassifierSimpleCV,
self).__init__(dyda_config_path=dyda_config_path)
self.set_param(self.class_name)
self.check_param_keys([
"bkg_ref_path", "re_width", "diff_thre", "pixel_num_min",
"kernel_size", "iter_num", "seg_margin"
])
self.orig_input_path = "BINARY_INPUT"
self.bkg_ref = image.read_img(self.param["bkg_ref_path"])
def test_main_process(self):
# pull test data from gitlab
config_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'abc67841cfb173c22cc13087fdd4d7b2/dyda.config'
dyda_config = lab_tools.pull_json_from_gitlab(config_url)
input_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'83db1facaef1bc83efc2b9ed25256ad1/park_input_list.json'
input_list = lab_tools.pull_json_from_gitlab(input_url)
output_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'775252d8456257f3364f788556bb553c/park_output_list.json'
output_list = lab_tools.pull_json_from_gitlab(output_url)
frame_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'aa4675a8057f3ae79bf5c0153cdfc69e/'\
'DeterminatorParkingLotStatus_frame_list.json'
frame_list = lab_tools.pull_json_from_gitlab(frame_url)
# initialization
determinator_ = DeterminatorParkingLotStatus(
dyda_config_path=dyda_config)
# run determinator
for i in range(len(output_list)):
# run determinator
determinator_.reset()
determinator_.input_data.append([image.read_img(frame_list[i])])
determinator_.input_data.append(input_list[i][0])
determinator_.input_data.append(input_list[i][1])
determinator_.input_data.append(input_list[i][2])
determinator_.run()
# compare results with reference
if not determinator_.results == []:
ref_data = output_list[i][0]
tar_data = determinator_.results[0]
report = dict_comparator.get_diff(ref_data, tar_data)
self.assertEqual(report['extra_field'], [])
self.assertEqual(report['missing_field'], [])
self.assertEqual(report['mismatch_val'], [])
def test_main_process(self):
# pull test data from gitlab
config_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'36356e92cad4a608d4c84bba769c0d53/'\
'dyda.config.DeterminatorCharacter'
dyda_config = lab_tools.pull_json_from_gitlab(config_url)
input_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'521feb8def6e63d6e187622b171b7233/input_list.json'
input_list = lab_tools.pull_json_from_gitlab(input_url)
output_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'9f25865b93a133bdec459ace16de8ccd/output_list.json'
output_list = lab_tools.pull_json_from_gitlab(output_url)
# initialization
determinator_ = DeterminatorCharacter(dyda_config_path=dyda_config)
# run determinator
for i in range(len(input_list)):
# run determinator
determinator_.reset()
determinator_.input_data.append(image.read_img(input_list[i]))
determinator_.input_data.append(
tools.parse_json(tools.replace_extension(
input_list[i], 'json')))
determinator_.run()
# compare results with reference
if not determinator_.results == []:
ref_data = output_list[i]
tar_data = determinator_.results
for j in range(len(ref_data)):
for k in range(len(ref_data[j])):
report = dict_comparator.get_diff(
ref_data[j][k], tar_data[j][k])
self.assertEqual(report['extra_field'], [])
self.assertEqual(report['missing_field'], [])
self.assertEqual(report['mismatch_val'], [])
def test_main_process(self):
# pull test data from gitlab
input_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'b7abe78697075dc2e47de5327c9f2de3/cv_classifier_input_list.json'
input_list = lab_tools.pull_json_from_gitlab(input_url)
output_url = 'https://gitlab.com/DT42/galaxy42/dt42-dyda/uploads/'\
'01fd55515c6373ced2541f0aec287f44/ref_bkg.jpg'
ref_output = lab_tools.pull_img_from_gitlab(output_url)
# initialization
learner_ = LearnerSimpleCV()
image_list = []
for i in range(len(input_list)):
image_list.append(image.read_img(input_list[i]))
# run classifier
learner_.reset()
learner_.input_data = image_list
learner_.run()
diff = tinycv.l1_norm_diff_cv2(learner_.output_data, ref_output)
self.assertEqual(sum(sum(diff)), 0.0)
def test_main_process():
dyda_config = {
'ClassifierTensorRT': {
'model_file':
'/home/nvidia/mobilenet_v1_1p0_224.engine',
'label_file':
'/home/nvidia/tf_to_trt_image_classification/data/imagenet_labels_1001.txt'
}
}
classifier_trt = ClassifierTensorRT(dyda_config_path=dyda_config)
input_path = '/home/nvidia/tf_to_trt_image_classification/data/images/golden_retriever.jpg'
classifier_trt.input_data = [image.read_img(input_path)]
# tr = tracker.SummaryTracker()
for i in range(int(sys.argv[1])):
print('Run #{}'.format(i))
s = time.time()
classifier_trt.run()
tar_data = classifier_trt.results[0]
print(tar_data)
print('classifier_trt.run() takes: {} ms'.format(
(time.time() - s) * 1000))