convert_labelImg_to_AutoML.py

import xmltodict
from pprint import pprint

with open('thumb0001.xml') as fd:
    doc = xmltodict.parse(fd.read())

def remap(x, in_min, in_max, out_min, out_max):
    return (x - in_min) * (out_max - out_min) / (in_max - in_min) + out_min

width = int(doc['annotation']['size']['width'])
height = int(doc['annotation']['size']['height'])
label = doc['annotation']['object']['name']
xmin = int(doc['annotation']['object']['bndbox']['xmin'])
ymin = int(doc['annotation']['object']['bndbox']['ymin'])
xmax = int(doc['annotation']['object']['bndbox']['xmax'])
ymax = int(doc['annotation']['object']['bndbox']['ymax'])


scaled_xmin = remap(xmin, 0, width,  0, 1)
scaled_ymin = remap(ymin, 0, height,  0, 1)

scaled_xmax = remap(xmax, 0, width,  0, 1)
scaled_ymax = remap(ymax, 0, height,  0, 1)

# print("Object: '{}' found @ ({},{})x({},{})".format(
#     label, scaled_xmin, scaled_ymin, scaled_xmax, scaled_ymax
# ))

print("TRAIN,gs://bucketname/imagename.jpg,{},{},{},,,{},{},,".format(
    label, scaled_xmin, scaled_ymin, scaled_xmax, scaled_ymax
))