def main(argv):
config_path = args.conf
num_anchors = args.anchors
with open(config_path) as config_buffer:
config = json.loads(config_buffer.read())
train_imgs, train_labels = parse_annotation(
config['train']['train_annot_folder'],
config['train']['train_image_folder'], config['model']['labels'])
grid_w = config['model']['input_size'] / 32
grid_h = config['model']['input_size'] / 32
# run k_mean to find the anchors
annotation_dims = []
for image in train_imgs:
cell_w = image['width'] / grid_w
cell_h = image['height'] / grid_h
for obj in image['object']:
relative_w = (float(obj['xmax']) - float(obj['xmin'])) / cell_w
relatice_h = (float(obj["ymax"]) - float(obj['ymin'])) / cell_h
annotation_dims.append(tuple(map(float, (relative_w, relatice_h))))
annotation_dims = np.array(annotation_dims)
centroids = run_kmeans(annotation_dims, num_anchors)
# write anchors to file
print('\naverage IOU for', num_anchors, 'anchors:',
'%0.2f' % avg_IOU(annotation_dims, centroids))
print_anchors(centroids)
def setValidationDataSet(self, data_path):
img_dir = os.path.join(data_path, 'train_97c/')
annotation_dir = os.path.join(data_path, 'annotation_97c')
img_dir += "/"
annotation_dir += "/"
self.valid_imgs, self.valid_labels = parse_annotation(
annotation_dir, img_dir, self.labels)
ocrdf = pd.read_csv(os.path.join(data_path, 'number_label.txt'))
self.ocr_valid_root_dir = data_path
self.ocr_valid_annotation = ocrdf.as_matrix()
self.ocr.setValidationData(self.ocr_valid_annotation,
self.ocr_valid_root_dir)
def test(self, data_path):
img_dir = os.path.join(data_path, 'train_97c/')
annotation_dir = os.path.join(data_path, 'annotation_97c')
img_dir += "/"
annotation_dir += "/"
test_imgs, test_labels = parse_annotation(annotation_dir, img_dir,
self.labels)
mAP = self.yolo.test(test_imgs)
print('Yolo Test Result, mAP:', mAP)
ocrdf = pd.read_csv(os.path.join(data_path, 'number_label.txt'))
test_annotation = ocrdf.as_matrix()
test_ds = self.ocr.getTestDataSet(test_annotation, data_path)
test_loss = self.ocr.test_model(test_ds)
print('OCR Test Loss:', test_loss)
return mAP[0], test_loss