-
Notifications
You must be signed in to change notification settings - Fork 4
/
dataset_cirr.py
175 lines (129 loc) · 5.76 KB
/
dataset_cirr.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
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
import os
import json as jsonmod
import torch
from dataset import MyDataset
from config import CIRR_IMAGE_DIR, CIRR_ANNOTATION_DIR
class CIRRDataset(MyDataset):
"""
CIRR (Composed Image Retrieval on Real-life Images), introduced in "Image
Retrieval on Real-Life Images With Pre-Trained Vision-and-Language Models";
Zheyuan Liu, Cristian Rodriguez-Opazo, Damien Teney, Stephen Gould;
Proceedings of the IEEE/CVF International Conference on Computer Vision
(ICCV), 2021, pp. 2125-2134
"""
def __init__(self, split, vocab, transform, what_elements='triplet', load_image_feature=0, ** kw):
MyDataset.__init__(self, split, CIRR_IMAGE_DIR, vocab, transform=transform, what_elements=what_elements, load_image_feature=load_image_feature, ** kw)
# NOTE: splits are refered as train/val/test but some filepaths contain resp. "train"/"dev"/"test1"
s = "test1" if split=="test" else split
# load the paths of the images involved in the split
self.image_name2path = self.load_file(os.path.join(CIRR_ANNOTATION_DIR, 'images_splits', f'split.rc2.{s}.json'))
self.image_id2name = list(self.image_name2path.keys())
# if necessary, load triplet annotations
if self.what_elements != "target":
self.annotations = self.load_file(os.path.join(CIRR_ANNOTATION_DIR, 'captions', f'cap.rc2.{s}.json'))
def __len__(self):
if self.what_elements=='target':
return len(self.image_id2name)
return len(self.annotations)
def load_file(self, f):
"""
If the input file is the "caption" annotation file, this returns a list
of dictionaries with the following format:
{"pairid": 12063,
"reference": "test1-147-1-img1",
"target_hard": "test1-83-0-img1",
"target_soft": {"test1-83-0-img1": 1.0},
"caption": "remove all but one dog and add a woman hugging it",
"img_set": {"id": 1,
"members": ["test1-147-1-img1",
"test1-1001-2-img0",
"test1-83-1-img1",
"test1-359-0-img1",
"test1-906-0-img1",
"test1-83-0-img1"],
"reference_rank": 3,
"target_rank": 4}
}
If the input file is about the images involved in the split, this
returns a list of image relative path
"""
with open(f, "r") as jsonfile:
ann = jsonmod.loads(jsonfile.read())
return ann
############################################################################
# *** GET ITEM METHODS
############################################################################
def get_triplet(self, index):
ann = self.annotations[index]
capts = ann['caption']
text, real_text = self.get_transformed_captions([capts])
path_src = self.image_name2path[ann['reference']][2:] # remove the "./" from the relative image pathname
img_src = self.get_transformed_image(path_src)
path_trg = self.image_name2path[ann['target_hard']][2:] # remove the "./" from the relative image pathname
img_trg = self.get_transformed_image(path_trg)
return img_src, text, img_trg, real_text, index
def get_query(self, index):
ann = self.annotations[index]
capts = ann['caption']
text, real_text = self.get_transformed_captions([capts])
path_src = self.image_name2path[ann['reference']][2:] # remove the "./" from the relative image pathname
img_src = self.get_transformed_image(path_src)
img_src_id = self.image_id2name.index(ann['reference'])
if self.split == "test":
img_trg_id = [None]
else:
img_trg_id = [self.image_id2name.index(ann['target_hard'])]
return img_src, text, img_src_id, img_trg_id, real_text, index
def get_target(self, index):
img_id = index
path_img = self.image_name2path[self.image_id2name[index]][2:] # remove the "./" from the relative image pathname
img = self.get_transformed_image(path_img)
return img, img_id, index
def get_subset(self, index):
"""
Get the ids of the images in the subset from wich the annotated
reference image and the target image are originated. These ids are
further used to compute the metrics R_subset@K defined for the CIRR
dataset cf. https://github.com/Cuberick-Orion/CIRR.
NOTE: the id of the reference image is not included in the subset !
"""
ann = self.annotations[index]
imgs_subset_ids = torch.tensor([self.image_id2name.index(im) \
for im in ann['img_set']['members']
if im != ann['reference']])
return imgs_subset_ids, index
def get_soft_targets(self, index):
"""
Get the ids of the soft-target images for the query indexed by
`index`, along with their qualification (1.0 when it's the actual
target image, 0.5 when no differences with the actual target image are
worth mentionning, -1.0 when the image is too different from the actual
target image). Currently, this is very specific to the CIRR dataset, and
is used at evaluation time to compute the recalls.
cf. https://github.com/Cuberick-Orion/CIRR.
"""
ann = self.annotations[index]
if self.split == "test":
softtrg_imgids_and_qualif = None
else:
softtrg_imgids_and_qualif = {self.image_id2name.index(im): qualif
for im, qualif in ann['target_soft'].items()}
return softtrg_imgids_and_qualif, index
def get_pair_ids(self):
"""
Returns each annotation's pair_id.
Necessary to produce prediction files, to be evaluated on the server.
"""
return [ann["pairid"] for ann in self.annotations]
############################################################################
# *** FORMATTING INFORMATION FOR VISUALIZATION PURPOSES
############################################################################
def get_triplet_info(self, index):
"""
Should return 3 strings:
- the text modifier
- an identification code (name, relative path...) for the reference image
- an identification code (name, relative path...) for the target image
"""
ann = self.annotations[index]
return ann["caption"], ann["reference"], ann["target_hard"]