How to make tfrecords from json annotated images

First I created csv files by using my own script.

import osimport globimport pandas as pdimport jsonimport pickledef json_to_csv():    path_to_json = 'images/train/'    json_files = [pos_json for pos_json in os.listdir(path_to_json) if pos_json.endswith('.json')]    path_to_jpeg = 'images/train/'    jpeg_files = [pos_jpeg for pos_jpeg in os.listdir(path_to_jpeg) if pos_jpeg.endswith('.jpeg')]    fjpeg=(list(reversed(jpeg_files)))    n=0    csv_list = []    labels=[]    for j in json_files:        data_file=open('images/train/{}'.format(j))           data = json.load(data_file)        width,height=data['display_width'],data['display_height']        for item in data["items"]:            box = item['bounding_box']            if item['upc']!='None':                name=item['upc']                labels.append(name)                xmin=box['left']                ymin=box['top']                xmax=box['right']                ymax=box['bottom']                value = (fjpeg[n],                         width,                         height,                         name,                         xmin,                         ymin,                         xmax,                         ymax                         )                csv_list.append(value)          n=n+1    column_name = ['filename', 'width', 'height', 'class', 'xmin', 'ymin', 'xmax', 'ymax']    csv_df = pd.DataFrame(csv_list, columns=column_name)    labels_train=list(set(labels))    with open("train_labels.txt", "wb") as fp:   #Pickling        pickle.dump(labels_train, fp)    return csv_dfdef main():    for directory in ['train']:        csv_df = json_to_csv()        csv_df.to_csv('data/{}_labels.csv'.format(directory), index=None)        print('Successfully converted json to csv.')main()

Then I use this script to create tfrecords.

We have some documentation on the subject.

Note that labelimg should produce outputs similar to the PASCAL VOC datasets we use, so those scripts may also be of use.