classes = [
'background', 'aeroplane', 'bicycle', 'bird', 'boat', 'bottle', 'bus',
'car', 'cat', 'chair', 'cow', 'diningtable', 'dog', 'horse', 'motorbike',
'person', 'pottedplant', 'sheep', 'sofa', 'train', 'tvmonitor'
]
train_dataset = BatchGenerator(
images_path='./Datasets/VOCdevkit/VOC2012/JPEGImages/',
include_classes='all',
box_output_format=['class_id', 'xmin', 'xmax', 'ymin', 'ymax'])
train_dataset.parse_xml(
annotations_path='./Datasets/VOCdevkit/VOC2012/Annotations/',
image_set_path='./Datasets/VOCdevkit/VOC2012/ImageSets/Main/',
image_set='train.txt',
classes=classes,
exclude_truncated=False,
exclude_difficult=False,
ret=False)
train_generator = train_dataset.generate(
batch_size=batch_size,
train=True,
ssd_box_encoder=ssd_box_encoder,
equalize=False,
brightness=(0.5, 2, 0.5),
flip=0.5,
translate=((0, 30), (0, 30), 0.5),
scale=(0.75, 1.2, 0.5),
random_crop=(
300, 300, 1,
# annotations_paths=[VOC_2007_annotations_path,
# VOC_2007_test_annotations_path,
# VOC_2012_annotations_path],
# image_set_paths=[VOC_2007_trainval_image_set_path,
# VOC_2007_test_image_set_path,
# VOC_2012_train_image_set_path],
# classes=classes,
# include_classes='all',
# exclude_truncated=False,
# exclude_difficult=False,
# ret=False)
#
val_dataset.parse_xml(images_paths=[road_test_images_path],
annotations_paths=[road_test_annotations_path],
image_set_paths=[road_test_image_set_path],
classes=classes,
include_classes='all',
exclude_truncated=False,
exclude_difficult=False,
ret=False)
# 3: Instantiate an encoder that can encode ground truth labels into the format needed by the SSD loss function.
# The encoder constructor needs the spatial dimensions of the model's predictor layers to create the anchor boxes.
predictor_sizes = [
model.get_layer('conv4_3_norm_mbox_conf').output_shape[1:3],
model.get_layer('fc7_mbox_conf').output_shape[1:3],
model.get_layer('conv6_2_mbox_conf').output_shape[1:3],
model.get_layer('conv7_2_mbox_conf').output_shape[1:3],
model.get_layer('conv8_2_mbox_conf').output_shape[1:3],
model.get_layer('conv9_2_mbox_conf').output_shape[1:3]
]
VOC_2007_test_images_dir = '../../datasets/VOCdevkit/VOC2007_Test/JPEGImages/'
VOC_2007_test_annotations_dir = '../../datasets/VOCdevkit/VOC2007_Test/Annotations/'
VOC_2007_test_image_set_filename = '../../datasets/VOCdevkit/VOC2007_Test/ImageSets/Main/test.txt'
# The XML parser needs to now what object class names to look for and in which order to map them to integers.
classes = [
'background', 'aeroplane', 'bicycle', 'bird', 'boat', 'bottle', 'bus',
'car', 'cat', 'chair', 'cow', 'diningtable', 'dog', 'horse', 'motorbike',
'person', 'pottedplant', 'sheep', 'sofa', 'train', 'tvmonitor'
]
dataset.parse_xml(images_dirs=[VOC_2007_test_images_dir],
image_set_filenames=[VOC_2007_test_image_set_filename],
annotations_dirs=[VOC_2007_test_annotations_dir],
classes=classes,
include_classes='all',
exclude_truncated=False,
exclude_difficult=True,
ret=False)
generator = dataset.generate(batch_size=1,
shuffle=True,
train=False,
returns={
'processed_images', 'filenames',
'inverse_transform', 'original_images',
'original_labels'
},
max_crop_and_resize=False,
random_pad_and_resize=False,
resize=(img_height, img_width))
PKLOT_images_dir = 'C:/Users/Damini/Documents/ECE477_ObjectDetection_Data_Storage/PKLot/PKLot'
# The directories that contain the annotations.
PKLOT_annotations_dir = 'C:/Users/Damini/Documents/ECE477_ObjectDetection_Data_Storage/PKLot2VOC_full_dataset/'
# Training dataset
PKLOT_train_image_set_filename = 'C:/Users/Damini/Documents/ECE477_ObjectDetection_Data_Storage/PKLot2VOC_full_dataset/ImageSets/Main/train.txt'
PKLOT_val_image_set_filename = 'C:/Users/Damini/Documents/ECE477_ObjectDetection_Data_Storage/PKLot2VOC_full_dataset/ImageSets/Main/valid.txt'
# The XML parser needs to now what object class names to look for and in which order to map them to integers.
classes = ['occupied', 'vaccant']
train_dataset.parse_xml(images_dirs=[PKLOT_images_dir],
image_set_filenames=[PKLOT_train_image_set_filename],
annotations_dirs=[PKLOT_annotations_dir],
classes=classes,
include_classes='all',
exclude_truncated=False,
exclude_difficult=False,
ret=False)
val_dataset.parse_xml(images_dirs=[PKLOT_images_dir],
image_set_filenames=[PKLOT_val_image_set_filename],
annotations_dirs=[PKLOT_annotations_dir],
classes=classes,
include_classes='all',
exclude_truncated=False,
exclude_difficult=True,
ret=False)
# The encoder constructor needs the spatial dimensions of the model's predictor layers to create the anchor boxes.
predictor_sizes = [
# ret=False)
# train_dataset.parse_xml(images_paths=[VOC_2007_images_path,VOC_2007_test_images_path],
# annotations_paths=[VOC_2007_annotations_path,VOC_2007_test_annotations_path],
# image_set_paths=[VOC_2007_train_image_set_path,VOC_2007_test_image_set_path],
# classes=classes,
# include_classes='all',
# exclude_truncated=False,
# exclude_difficult=False,
# ret=False)
train_dataset.parse_xml(
images_paths=[apple_images, apple_images],
annotations_paths=[apple_annotations, apple_annotations],
image_set_paths=[apple_train_imageset, apple_val_imageset],
classes=classes,
include_classes='all',
exclude_truncated=False,
exclude_difficult=False,
ret=False)
train_generator = train_dataset.generate(
batch_size=batch_size,
train=True,
ssd_box_encoder=ssd_box_encoder,
equalize=False,
brightness=(0.5, 2, 0.5),
flip=0.5,
translate=False,
scale=False,
max_crop_and_resize=(