def eval_ufo():
os.environ['CUDA_VISIBLE_DEVICES'] = '0,1,2,3'
is_tf1 = True
on_test = False
main_dir = "/home/jsteeg/MRP_Implementations/tf1/out"
model_path = os.path.join(main_dir, "exported_model")
output_dir = os.path.join(
main_dir, "validate_results" + ("" if on_test else "_train"))
set = "test" if on_test else "train"
dataset_file = f"/home/jsteeg/ufo_data/yolo_no_crop_vc/{set}.txt"
label_map = "/home/jsteeg/ufo_data/yolo_no_crop_vc/label_map.txt"
save_output = True
id_list = None
threshold = 0.05
if is_tf1:
detector = DetectorTF1(model_path,
label_map,
class_id=id_list,
threshold=threshold)
else:
detector = DetectorTF2(model_path,
label_map,
class_id=id_list,
threshold=threshold)
validate(detector=detector,
dataset_file=dataset_file,
output_dir=output_dir,
save_output=save_output)
def __init__(self):
self.corner_detection_model = Detector(
path_to_model='./models/identity_corner/model.tflite',
path_to_labels='./models/identity_corner/label_map.pbtxt',
nms_threshold=0.2,
score_threshold=0.3)
self.text_detection_model = Detector(
path_to_model='./models/identity_card/model.tflite',
path_to_labels='./models/identity_card/label_map.pbtxt',
nms_threshold=0.2,
score_threshold=0.2)
self.text_detection_discharge = DetectorTF2(
path_to_model='./models/discharge_record',
path_to_labels='./models/discharge_record/label_map.pbtxt',
nms_threshold=0.33,
score_threshold=0.33)
self.text_recognition_model = TextRecognition(
path_to_checkpoint='./models/text_recogintion/transformerocr.pth')
default='models/mscoco_label_map.pbtxt')
parser.add_argument('--class_ids', help='id of classes to detect, expects string with ids delimited by ","',
type=str, default=None) # example input "1,3" to detect person and car
parser.add_argument('--threshold', help='Detection Threshold', type=float, default=0.4)
parser.add_argument('--images_dir', help='Directory to input images)', default='data/samples/images/')
parser.add_argument('--video_path', help='Path to input video)', default='data/samples/pedestrian_test.mp4')
parser.add_argument('--output_directory', help='Path to output images and video', default='data/samples/output')
parser.add_argument('--video_input', help='Flag for video input, default: False', action='store_true') # default is false
parser.add_argument('--save_output', help='Flag for save images and video with detections visualized, default: False',
action='store_true') # default is false
args = parser.parse_args()
id_list = None
if args.class_ids is not None:
id_list = [int(item) for item in args.class_ids.split(',')]
if args.save_output:
if not os.path.exists(args.output_directory):
os.makedirs(args.output_directory)
# instance of the class DetectorTF2
detector = DetectorTF2(args.model_path, args.path_to_labelmap, class_id=id_list, threshold=args.threshold)
if args.video_input:
DetectFromVideo(detector, args.video_path, save_output=args.save_output, output_dir=args.output_directory)
else:
DetectImagesFromFolder(detector, args.images_dir, save_output=args.save_output, output_dir=args.output_directory)
print("Done ...")
#cv2.destroyAllWindows()