-
Notifications
You must be signed in to change notification settings - Fork 0
/
generate_coco_json_5_classes_Malaga_5m.py
117 lines (102 loc) · 6.4 KB
/
generate_coco_json_5_classes_Malaga_5m.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
#version log
# adding list_of_image_numbers for create_annotations function
# generating random list of validation images
import datetime
import json
global date
date = str(datetime.datetime.now())
print(date)
import random
seed = 30
#master_json_dict = {} #The final dictionary that will be dumped into the json
info__dict = {
"year": 2019,
"version": 1,
"description": "Coco style dataset with five classes for object detection test in Malaga",
"contributor": "Sai Abinesh",
"url": "10.5281/zenodo.1403708",
"date_created": date,
}
#Changing id 10 from traffic light to gas cylinder
categories_list = [{"supercategory": "animal", "id": 1, "name": "dead animal"}, {"supercategory": "animal", "id": 2, "name": "dead bird"}, {"supercategory": "outdoor", "id": 3, "name": "gas cylinder"},{"supercategory": "accessory", "id": 4, "name": "suitcase"},{"supercategory": "accessory", "id": 5, "name": "backpack"}]
licenses_list = [{"url": "http:\/\/creativecommons.org\/licenses\/by-nc-sa\/2.0\/", "id": 1, "name": "Attribution-NonCommercial-ShareAlike License"}, {"url": "http:\/\/creativecommons.org\/licenses\/by-nc\/2.0\/", "id": 2, "name": "Attribution-NonCommercial License"}, {"url": "http:\/\/creativecommons.org\/licenses\/by-nc-nd\/2.0\/", "id": 3, "name": "Attribution-NonCommercial-NoDerivs License"}, {"url": "http:\/\/creativecommons.org\/licenses\/by\/2.0\/", "id": 4, "name": "Attribution License"}, {"url": "http:\/\/creativecommons.org\/licenses\/by-sa\/2.0\/", "id": 5, "name": "Attribution-ShareAlike License"}, {"url": "http:\/\/creativecommons.org\/licenses\/by-nd\/2.0\/", "id": 6, "name": "Attribution-NoDerivs License"}, {"url": "http:\/\/flickr.com\/commons\/usage\/", "id": 7, "name": "No known copyright restrictions"}, {"url": "http:\/\/www.usa.gov\/copyright.shtml", "id": 8, "name": "United States Government Work"}]
#Creating the master json dictionary using the prepopulated fields above
master_json_dict = {"info": info__dict, "images": [], "annotations": [], "licenses": licenses_list, "categories": categories_list }
# print("categories list for id 10: ", categories_list[9])
# print(master_json_dict)
# The following function outputs a python dictionary that can be written into a file as a json. Right now bounding boxes are not written
def create_annotations(category_id_list, folder_path, master_json_dict, path_binary_list_of_flags, list_of_image_numbers, file_path_for_json):
global date
# Writing the basic image info for each image
annotation_count =1
image_count =1
for image_number in list_of_image_numbers:
image_temp = {
"id": image_number,
"width": 1024,
"height": 1024,
"file_name": folder_path+"/image_"+str(image_number)+"_raw.png",
"license": 1,
"flickr_url": "none",
"coco_url": "none",
"date_captured": date,
}
list_of_annotations_current_image = []
for i in range(0,len(category_id_list)):
current_category_id = category_id_list[i][0] #The first element in each list is the category id
current_category_name = category_id_list[i][1] #The second element is the category name
number_of_instances_in_category = category_id_list[i][2] #The third element is the number of instances in that category
for instance_number in range(1,number_of_instances_in_category+1):
#Reading the text file containing the binary flags of 1s and 0s and storing it into a list. The text file is of the format ex. "list_gas cylinder_5"
binary_flags_object_present = open(path_binary_list_of_flags+"/list_"+current_category_name+"_"+str(instance_number)+".txt").read().splitlines()
#Converting the elements of the list into int
binary_flags_object_present = [int(item) for item in binary_flags_object_present]
#if not ((len(binary_flags_object_present)) == 500):
# print(len(binary_flags_object_present))
# print((path_binary_list_of_flags+"/list_"+current_category_name+"_"+str(instance_number)+".txt"))
if binary_flags_object_present[image_number-1]: # image_number-1 since image_number is indexed from 1, and binary flags are python indexed from 0
annotations_temp = {
"id": annotation_count,
"image_id": image_number,
"category_id": current_category_id,
"segmentation": [],
"area": 0,
"bbox": [], #For now ignoring this as bounding boxes will be automatically calculated from binary masks which are provided as numpy array from binary B/W style images
"iscrowd": 0,
"instance_number": instance_number,
#"annotation_id":annotation_count,
}
annotation_count = annotation_count + 1
list_of_annotations_current_image.append(annotations_temp)
master_json_dict["images"].append(image_temp)
master_json_dict["annotations"] = [*master_json_dict["annotations"], *list_of_annotations_current_image]
# For each image writing annotations of objects so that the corresponding masks can be called during the finetuning process
# The category id, category name, and the number of instances in each category are passed in as a list of lists as [[category_id, category name, number_of_instances],...]
image_count = image_count+1
if (image_count%50)==0:
print("Images done: ",image_count)
# print(image_number)
with open(file_path_for_json, 'w') as data_file:
json.dump(master_json_dict,data_file)
# seed = 30
# random.seed(seed)
# validation_set = random.sample(range(1,401),40)
seed = 30
random.seed(seed)
validation_set = random.sample(range(401,501),10)
# training_set = [i for i in range(1,401) if i not in validation_set]
training_set = [i for i in range(401,501) if i not in validation_set]
# print(training_set)
# print(len(training_set))
# exit()
folder_path = "/ichec/home/users/saiabinesh/experiments/maskrcnn/images/Images_master_v2/Images_master_v2"
category_id_list = [[1,"dead animal", 30],[2, "dead bird", 30],[3,"gas cylinder",57],[4,"suitcase",67],[5,"backpack",116]]
path_binary_list_of_flags = "D:/AirSim/New/AirSim/PythonClient/computer_vision"
train_path = "D:/instances_train2014.json"
print("Creating training annotations")
create_annotations(category_id_list, folder_path, master_json_dict, path_binary_list_of_flags,training_set, train_path)
#resetting master_json_dict before generating the next json
master_json_dict = {"info": info__dict, "images": [], "annotations": [], "licenses": licenses_list, "categories": categories_list }
print("Now Creating val annotations")
val_path = "D:/instances_val2014.json"
create_annotations(category_id_list, folder_path, master_json_dict, path_binary_list_of_flags,validation_set, val_path)