Datasets:

Size Categories:
n<1K
ArXiv:
Tags:
License:
image
imagewidth (px)
113
5.62k
label
imagewidth (px)
113
5.62k
classes_on_image
sequencelengths
2
12
id
int64
0
4.98k
[ 0, 48, 66, 90 ]
0
[ 0, 52, 70, 82, 93 ]
1
[ 0, 58, 67 ]
2
[ 0, 48, 84, 85, 87 ]
3
[ 0, 46, 84, 87, 96, 101 ]
4
[ 0, 48, 73, 89 ]
5
[ 0, 49, 52, 59, 70, 84 ]
6
[ 0, 49, 66, 70 ]
7
[ 0, 9, 24, 30, 58 ]
8
[ 0, 47, 84, 95 ]
9
[ 0, 5, 9 ]
10
[ 0, 3, 8 ]
11
[ 0, 8, 73 ]
12
[ 0, 46, 52, 58, 73, 80 ]
13
[ 0, 46, 71, 84 ]
14
[ 0, 54, 84, 87, 96 ]
15
[ 0, 37, 54, 93 ]
16
[ 0, 3, 46, 67 ]
17
[ 0, 55, 58, 73, 76 ]
18
[ 0, 14, 37, 48, 52, 94, 103 ]
19
[ 0, 8, 72, 81, 82, 84, 87, 94 ]
20
[ 0, 37, 48 ]
21
[ 0, 14, 25, 29, 44 ]
22
[ 0, 8, 84, 88 ]
23
[ 0, 46, 59, 67, 101 ]
24
[ 0, 47, 59, 85 ]
25
[ 0, 46, 96 ]
26
[ 0, 48, 66, 96 ]
27
[ 0, 48, 70, 84, 85, 87 ]
28
[ 0, 46, 58, 70, 84, 87 ]
29
[ 0, 32, 44, 89, 93 ]
30
[ 0, 58, 73, 76, 80 ]
31
[ 0, 47, 52, 72 ]
32
[ 0, 46, 52, 76, 77 ]
33
[ 0, 48, 65, 73, 87 ]
34
[ 0, 48, 70, 84, 89, 96 ]
35
[ 0, 5, 57 ]
36
[ 0, 48, 52 ]
37
[ 0, 15, 30, 37, 89 ]
38
[ 0, 58, 73, 89 ]
39
[ 0, 48, 84, 87 ]
40
[ 0, 37, 54, 85 ]
41
[ 0, 50, 52, 89 ]
42
[ 0, 8, 44 ]
43
[ 0, 37, 42, 44, 54, 73, 80, 89 ]
44
[ 0, 48, 58, 59, 73, 93 ]
45
[ 0, 47, 56, 65, 73, 96 ]
46
[ 0, 46, 70, 84, 89 ]
47
[ 0, 72, 82, 84, 87, 96 ]
48
[ 0, 47, 67, 84 ]
49
[ 0, 52, 67, 73, 89 ]
50
[ 0, 61, 73, 93, 94 ]
51
[ 0, 3, 47, 58, 73 ]
52
[ 0, 5, 8, 84 ]
53
[ 0, 30, 58 ]
54
[ 0, 46, 84, 87 ]
55
[ 0, 8, 30, 52, 58, 89 ]
56
[ 0, 28, 73, 80, 89 ]
57
[ 0, 8, 52, 58 ]
58
[ 0, 3, 28, 52, 73 ]
59
[ 0, 37, 87 ]
60
[ 0, 52, 58 ]
61
[ 0, 14, 44, 89 ]
62
[ 0, 52, 62 ]
63
[ 0, 14, 37 ]
64
[ 0, 48, 50, 66, 87, 95, 101 ]
65
[ 0, 24, 54, 72, 82, 84, 87, 96 ]
66
[ 0, 9, 14, 52, 58, 73, 93 ]
67
[ 0, 30, 58 ]
68
[ 0, 3, 46, 58, 59, 80 ]
69
[ 0, 3, 52, 59, 67, 96 ]
70
[ 0, 37, 52, 54, 56, 66, 85 ]
71
[ 0, 32, 58 ]
72
[ 0, 48, 49, 52, 66, 70, 103 ]
73
[ 0, 4, 10, 24 ]
74
[ 0, 37, 48, 59, 89 ]
75
[ 0, 37, 48, 66, 72, 84, 87 ]
76
[ 0, 24, 66, 73, 80, 102 ]
77
[ 0, 30, 58 ]
78
[ 0, 9, 41, 58, 73 ]
79
[ 0, 67, 73 ]
80
[ 0, 9, 48, 66, 87 ]
81
[ 0, 52, 76, 82, 87, 94 ]
82
[ 0, 28, 54, 58, 73, 80, 93 ]
83
[ 0, 47, 71, 77, 100 ]
84
[ 0, 48, 50, 72, 89, 101 ]
85
[ 0, 29, 30, 40, 52 ]
86
[ 0, 58, 94 ]
87
[ 0, 47, 66, 95 ]
88
[ 0, 10, 73, 94 ]
89
[ 0, 58, 59, 70 ]
90
[ 0, 48, 66, 84 ]
91
[ 0, 10, 30, 89 ]
92
[ 0, 3, 46, 52, 58, 80 ]
93
[ 0, 24, 44, 52, 67, 89 ]
94
[ 0, 46, 70, 85 ]
95
[ 0, 10, 52 ]
96
[ 0, 54, 70, 84, 96 ]
97
[ 0, 3, 8, 48, 93 ]
98
[ 0, 49, 67, 73, 93 ]
99

Dataset Card for FoodSeg103

Dataset Summary

FoodSeg103 is a large-scale benchmark for food image segmentation. It contains 103 food categories and 7118 images with ingredient level pixel-wise annotations. The dataset is a curated sample from Recipe1M and annotated and refined by human annotators. The dataset is split into 2 subsets: training set, validation set. The training set contains 4983 images and the validation set contains 2135 images.

Supported Tasks and Leaderboards

No leaderboard is available for this dataset at the moment.

Dataset Structure

Data categories

id ingridient
0 background
1 candy
2 egg tart
3 french fries
4 chocolate
5 biscuit
6 popcorn
7 pudding
8 ice cream
9 cheese butter
10 cake
11 wine
12 milkshake
13 coffee
14 juice
15 milk
16 tea
17 almond
18 red beans
19 cashew
20 dried cranberries
21 soy
22 walnut
23 peanut
24 egg
25 apple
26 date
27 apricot
28 avocado
29 banana
30 strawberry
31 cherry
32 blueberry
33 raspberry
34 mango
35 olives
36 peach
37 lemon
38 pear
39 fig
40 pineapple
41 grape
42 kiwi
43 melon
44 orange
45 watermelon
46 steak
47 pork
48 chicken duck
49 sausage
50 fried meat
51 lamb
52 sauce
53 crab
54 fish
55 shellfish
56 shrimp
57 soup
58 bread
59 corn
60 hamburg
61 pizza
62 hanamaki baozi
63 wonton dumplings
64 pasta
65 noodles
66 rice
67 pie
68 tofu
69 eggplant
70 potato
71 garlic
72 cauliflower
73 tomato
74 kelp
75 seaweed
76 spring onion
77 rape
78 ginger
79 okra
80 lettuce
81 pumpkin
82 cucumber
83 white radish
84 carrot
85 asparagus
86 bamboo shoots
87 broccoli
88 celery stick
89 cilantro mint
90 snow peas
91 cabbage
92 bean sprouts
93 onion
94 pepper
95 green beans
96 French beans
97 king oyster mushroom
98 shiitake
99 enoki mushroom
100 oyster mushroom
101 white button mushroom
102 salad
103 other ingredients

Data Splits

This dataset only contains two splits. A training split and a validation split with 4983 and 2135 images respectively.

Dataset Creation

Curation Rationale

Select images from a large-scale recipe dataset and annotate them with pixel-wise segmentation masks.

Source Data

The dataset is a curated sample from Recipe1M.

Initial Data Collection and Normalization

After selecting the source of the data two more steps were added before image selection.

  1. Recipe1M contains 1.5k ingredient categoris, but only the top 124 categories were selected + a 'other' category (further became 103).
  2. Images should contain between 2 and 16 ingredients.
  3. Ingredients should be visible and easy to annotate.

Which then resulted in 7118 images.

Annotations

Annotation process

Third party annotators were hired to annotate the images respecting the following guidelines:

  1. Tag ingredients with appropriate categories.
  2. Draw pixel-wise masks for each ingredient.
  3. Ignore tiny regions (even if contains ingredients) with area covering less than 5% of the image.

Refinement process

The refinement process implemented the following steps:

  1. Correct mislabelled ingredients.
  2. Deleting unpopular categories that are assigned to less than 5 images (resulting in 103 categories in the final dataset).
  3. Merging visually similar ingredient categories (e.g. orange and citrus)

Who are the annotators?

A third party company that was not mentioned in the paper.

Additional Information

Dataset Curators

Authors of the paper A Large-Scale Benchmark for Food Image Segmentation.

Licensing Information

Apache 2.0 license.

Citation Information

@inproceedings{wu2021foodseg,
    title={A Large-Scale Benchmark for Food Image Segmentation},
    author={Wu, Xiongwei and Fu, Xin and Liu, Ying and Lim, Ee-Peng and Hoi, Steven CH and Sun, Qianru},
    booktitle={Proceedings of ACM international conference on Multimedia},
    year={2021}
}
Downloads last month
3,058
Edit dataset card

Models trained or fine-tuned on EduardoPacheco/FoodSeg103