class FCOSEvaluator(BaseEvaluator):
def __init__(self, iou_thresh, detector_name, weights, config=None):
self.THRESHOLDS_FOR_CLASSES = [
0.3, 0.4928510785102844, 0.5040897727012634,
0.4912887513637543, 0.5016880631446838, 0.5278812646865845,
0.5351834893226624, 0.5003424882888794, 0.4955945909023285,
0.43564629554748535, 0.6089804172515869, 0.666087806224823,
0.5932040214538574, 0.48406165838241577, 0.4062422513961792,
0.5571075081825256, 0.5671307444572449, 0.5268378257751465,
0.5112953186035156, 0.4647842049598694, 0.5324517488479614,
0.5795850157737732, 0.5152440071105957, 0.5280804634094238,
0.4791383445262909, 0.5261335372924805, 0.4906163215637207,
0.523737907409668, 0.47027698159217834, 0.5103300213813782,
0.4645252823829651, 0.5384289026260376, 0.47796186804771423,
0.4403403103351593, 0.5101461410522461, 0.5535093545913696,
0.48472103476524353, 0.5006796717643738, 0.5485560894012451,
0.4863888621330261, 0.5061569809913635, 0.5235867500305176,
0.4745445251464844, 0.4652363359928131, 0.4162440598011017,
0.5252017974853516, 0.42710989713668823, 0.4550687372684479,
0.4943239390850067, 0.4810051918029785, 0.47629663348197937,
0.46629616618156433, 0.4662836790084839, 0.4854755401611328,
0.4156557023525238, 0.4763634502887726, 0.4724511504173279,
0.4915047585964203, 0.5006274580955505, 0.5124194622039795,
0.47004589438438416, 0.5374764204025269, 0.5876904129981995,
0.49395060539245605, 0.5102297067642212, 0.46571290493011475,
0.5164387822151184, 0.540651798248291, 0.5323763489723206,
0.5048757195472717, 0.5302401781082153, 0.48333442211151123,
0.5109739303588867, 0.4077408015727997, 0.5764586925506592,
0.5109297037124634, 0.4685552418231964, 0.5148998498916626,
0.4224434792995453, 0.4998510777950287
]
super(FCOSEvaluator, self).__init__(iou_thresh, detector_name, weights, config=config)
def read_json(self, filename):
"""
Decodes a JSON file & returns its content.
Raises:
FileNotFoundError: file not found
ValueError: failed to decode the JSON file
TypeError: the type of decoded content differs from the expected (list of dictionaries)
:param filename: [str] name of the JSON file
:return: [list] list of the annotations
"""
if not os.path.exists(filename):
raise FileNotFoundError("File %s not found." % filename)
try:
with open(filename, 'r') as _f:
_data = json.load(_f)
except json.JSONDecodeError:
raise ValueError(f"Failed to decode {filename}.")
if not isinstance(_data, list):
raise TypeError(f"Decoded content is {type(_data)}. Expected list.")
if len(_data) > 0 and not isinstance(_data[0], dict):
raise TypeError(f"Decoded content is {type(_data[0])}. Expected dict.")
return _data
# get the ground-truth bbox coordinates from the json annotation files
def load_ground_truth_coordinates(self, filename, view=None): # filename format is 0000XXXX.png, XXXX is a timestamp
if view == None:
raise TypeError("view cannot be None")
annotations = self.read_json(filename)
coordinates = list()
for annotation in annotations:
bbox = annotation['views'][view-1] # 0 - 6, representing C1 - C 7
xmin = bbox['xmin']
ymin = bbox['ymin']
xmax = bbox['xmax']
ymax = bbox['ymax']
if (xmin, ymin, xmax, ymax) == (-1, -1, -1, -1): # person not present
continue
coordinate = list()
coordinate.append((xmin, ymin))
coordinate.append((xmax, ymax))
coordinates.append(coordinate)
return coordinates
def prepare_detector(self):
# Load model
cfg.merge_from_file(self.config)
cfg.merge_from_list(list())
cfg.MODEL.WEIGHT = self.weights
cfg.freeze()
self.detector = COCODemo(
cfg,
confidence_thresholds_for_classes=self.THRESHOLDS_FOR_CLASSES,
min_image_size=800)
def load_detections_and_confidence(self, image):
#self.prepare_detector()
detections, confidence = self.detector.get_person_detections(image)
return detections, confidence
def evaluate_one_set(self, view):
# for view in range(7): # examine 1 video at a time
for time_stamp in range(0, 1996, 5): # examine each frame
if int(time_stamp / 10) == 0: # single digit 000X
stamp = "000" + str(time_stamp)
elif int(time_stamp / 100) == 0: # two digits 00XX
stamp = "00" + str(time_stamp)
elif int(time_stamp / 1000) == 0: # three digits 0XXX
stamp = "0" + str(time_stamp)
else: # XXXX
stamp = str(time_stamp)
print("Processing C" + str(view) + "0000" + stamp + ".png")
img = cv2.imread("../../Wildtrack_dataset/Image_subsets/C" + str(view) + "/0000" + stamp + ".png")
detections, confidence = self.load_detections_and_confidence(img)
ground_truth_coordinates = self.load_ground_truth_coordinates("../../Wildtrack_dataset/annotations_positions/0000" + stamp + ".json", view=view)
self.classify_detections(detections, confidence, ground_truth_coordinates, "C"+str(view)+"0000"+stamp+".png")
#self.calculate_final_stats()
def evaluate(self):
#x = threading.Thread(target=thread_function, args=(1,))
self.prepare_detector()
threads = list()
for view in range(1, 8): # examine 1 video at a time
#print("view:", view)
#t = threading.Thread(target=self.evaluate_one_set, args=(view,))
#t.start()
#threads.append(t)
self.evaluate_one_set(view)
#for thread in threads:
#thread.join()
self.calculate_final_stats()
for view in range(7): # examine 1 video at a time
for time_stamp in range(0, 1996, 5): # examine each frame
if int(time_stamp / 10) == 0: # single digit 000X
stamp = "000" + str(time_stamp)
elif int(time_stamp / 100) == 0: # two digits 00XX
stamp = "00" + str(time_stamp)
elif int(time_stamp / 1000) == 0: # three digits 0XXX
stamp = "0" + str(time_stamp)
else: # XXXX
stamp = str(time_stamp)
print("Processing C" + str(view + 1) + "0000" + stamp + ".png")
# prepare object that handles inference plus adds predictions on top of image
img = cv2.imread("../../Wildtrack_dataset/Image_subsets/C" +
str(view + 1) + "/0000" + stamp + ".png")
detections, confidence = coco_demo.get_person_detections(img)
ground_truth_coordinates = load_ground_truth_coordinates(
view, "../../Wildtrack_dataset/annotations_positions/0000" +
stamp + ".json")
num_ground_truth = len(ground_truth_coordinates)
total_gt_boxes += num_ground_truth # add to number of ground-truth bboxes
# create dataframe for each image, storing information regarding each prediction it contains
#Later appended to the "overall" dataframe
image_dataframe = pd.DataFrame(columns=("ImageID", "coordinate",
"confidence", "TP", "FP"))
image_dataframe["ImageID"] = [
"C" + str(view + 1) + "0000" + stamp
def main():
parser = argparse.ArgumentParser(
description="PyTorch Object Detection Webcam Demo")
parser.add_argument(
"--config-file",
default="configs/fcos/fcos_imprv_R_50_FPN_1x.yaml",
metavar="FILE",
help="path to config file",
)
parser.add_argument(
"--weights",
default="weights/FCOS_imprv_R_50_FPN_1x.pth",
metavar="FILE",
help="path to the trained model",
)
parser.add_argument(
"--images-dir",
default="demo/images",
metavar="DIR",
help="path to demo images directory",
)
parser.add_argument(
"--min-image-size",
type=int,
default=800,
help="Smallest size of the image to feed to the model. "
"Model was trained with 800, which gives best results",
)
parser.add_argument(
"opts",
help="Modify model config options using the command-line",
default=None,
nargs=argparse.REMAINDER,
)
args = parser.parse_args()
print(args.opts)
# load config from file and command-line arguments
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.MODEL.WEIGHT = args.weights
cfg.freeze()
# The following per-class thresholds are computed by maximizing
# per-class f-measure in their precision-recall curve.
# Please see compute_thresholds_for_classes() in coco_eval.py for details.
thresholds_for_classes = [
0.3, 0.4928510785102844, 0.5040897727012634, 0.4912887513637543,
0.5016880631446838, 0.5278812646865845, 0.5351834893226624,
0.5003424882888794, 0.4955945909023285, 0.43564629554748535,
0.6089804172515869, 0.666087806224823, 0.5932040214538574,
0.48406165838241577, 0.4062422513961792, 0.5571075081825256,
0.5671307444572449, 0.5268378257751465, 0.5112953186035156,
0.4647842049598694, 0.5324517488479614, 0.5795850157737732,
0.5152440071105957, 0.5280804634094238, 0.4791383445262909,
0.5261335372924805, 0.4906163215637207, 0.523737907409668,
0.47027698159217834, 0.5103300213813782, 0.4645252823829651,
0.5384289026260376, 0.47796186804771423, 0.4403403103351593,
0.5101461410522461, 0.5535093545913696, 0.48472103476524353,
0.5006796717643738, 0.5485560894012451, 0.4863888621330261,
0.5061569809913635, 0.5235867500305176, 0.4745445251464844,
0.4652363359928131, 0.4162440598011017, 0.5252017974853516,
0.42710989713668823, 0.4550687372684479, 0.4943239390850067,
0.4810051918029785, 0.47629663348197937, 0.46629616618156433,
0.4662836790084839, 0.4854755401611328, 0.4156557023525238,
0.4763634502887726, 0.4724511504173279, 0.4915047585964203,
0.5006274580955505, 0.5124194622039795, 0.47004589438438416,
0.5374764204025269, 0.5876904129981995, 0.49395060539245605,
0.5102297067642212, 0.46571290493011475, 0.5164387822151184,
0.540651798248291, 0.5323763489723206, 0.5048757195472717,
0.5302401781082153, 0.48333442211151123, 0.5109739303588867,
0.4077408015727997, 0.5764586925506592, 0.5109297037124634,
0.4685552418231964, 0.5148998498916626, 0.4224434792995453,
0.4998510777950287
]
demo_im_names = os.listdir(args.images_dir)
# prepare object that handles inference plus adds predictions on top of image
coco_demo = COCODemo(
cfg,
confidence_thresholds_for_classes=thresholds_for_classes,
min_image_size=args.min_image_size)
print("starting...")
for im_name in demo_im_names:
print(im_name)
img = cv2.imread(os.path.join(args.images_dir, im_name))
if img is None:
continue
start_time = time.time()
composite = coco_demo.run_on_opencv_image(img)
coordinates, scores = coco_demo.get_person_detections(img)
print("coordinates:", coordinates, "scores:", scores)
print("{}\tinference time: {:.2f}s".format(im_name,
time.time() - start_time))
#cv2.imshow(im_name, composite)
print("saving...")
cv2.imwrite("demo_output/" + im_name, composite)
