convert_tf_to_coco.py

"""
Use this to convert csv -> df -> coco
"""

import numpy as np
import json
import pandas as pd
import os

# use this method to convert csv to json

def convert_to_coco(data, save_json_path):
    #path = 'tensorflow2csv.csv'
    #data = pd.read_csv(path)

    images = []
    categories = []
    annotations = []

    category = {}
    category["supercategory"] = 'none'
    category["id"] = 0
    category["name"] = 'None'
    categories.append(category)

    data['fileid'] = data['filename'].astype('category').cat.codes
    data['categoryid']= pd.Categorical(data['class'],ordered= True).codes
    data['categoryid'] = data['categoryid']+1
    data['annid'] = data.index

    def image(row):
        image = {}
        image["height"] = row.height
        image["width"] = row.width
        image["id"] = row.fileid
        image["file_name"] = row.filename
        return image

    def category(row):
        category = {}
        category["supercategory"] = 'None'
        category["id"] = row.categoryid
        category["name"] = row[2]
        return category

    def annotation(row):
        annotation = {}
        area = (row.xmax -row.xmin)*(row.ymax - row.ymin)
        annotation["segmentation"] = []
        annotation["iscrowd"] = 0
        annotation["area"] = area
        annotation["image_id"] = row.fileid

        annotation["bbox"] = [row.xmin, row.ymin, row.xmax -row.xmin,row.ymax-row.ymin ]

        annotation["category_id"] = row.categoryid
        annotation["id"] = row.annid
        return annotation

    for row in data.itertuples():
        annotations.append(annotation(row))

    imagedf = data.drop_duplicates(subset=['fileid']).sort_values(by='fileid')
    for row in imagedf.itertuples():
        images.append(image(row))

    catdf = data.drop_duplicates(subset=['categoryid']).sort_values(by='categoryid')
    for row in catdf.itertuples():
        categories.append(category(row))

    data_coco = {}
    data_coco["images"] = images
    data_coco["categories"] = categories
    data_coco["annotations"] = annotations


    json.dump(data_coco, open(save_json_path, "w"), indent=4)

def main():
    #dddd
    root = os.getcwd()
    print(root)

    train_csv_path = r'C:\Users\endle\Desktop\pytorch-retinanet\train_labels  - org.csv'
    test_csv_path = r'C:\Users\endle\Desktop\pytorch-retinanet\test_labels - org.csv'

    df_train = pd.read_csv(train_csv_path)
    df_test = pd.read_csv(test_csv_path)

    save_json_path_train = r'C:\Users\endle\Desktop\pytorch-retinanet\train_new.json'
    save_json_path_val = r'C:\Users\endle\Desktop\pytorch-retinanet\val_new.json'

    convert_to_coco(df_train, save_json_path_train)
    convert_to_coco(df_test, save_json_path_val)


if __name__ == "__main__":
    main()