Hugging Face's logo

Datasets:
bazyl
/
GTSRB

Tasks:
Image Classification
Sub-tasks: multi-label-image-classification
Size Categories: 10K<n<100K
Language Creators: found
Annotations Creators: crowdsourced
Source Datasets: original
Tags:
License:
Dataset card Viewer Files Community
1
Dataset Viewer
Auto-converted to Parquet
Viewer
Width
uint16
25
243
Height
uint8
25
225
Roi.X1
uint8
0
20
Roi.Y1
uint8
5
20
Roi.X2
uint8
20
223
Roi.Y2
uint8
20
205
ClassId
class label
0
42
Path
image
27
26
5
5
22
20
2020
28
27
5
6
23
22
2020
29
26
6
5
24
21
2020
28
27
5
6
23
22
2020
28
26
5
5
23
21
2020
31
27
6
5
26
22
2020
31
28
6
6
26
23
2020
31
28
6
6
26
23
2020
31
29
5
6
26
24
2020
34
32
6
6
29
26
2020
36
33
5
6
31
28
2020
37
34
5
6
32
29
2020
38
34
5
6
32
29
2020
40
34
6
6
34
29
2020
39
34
5
5
34
29
2020
42
36
6
5
37
31
2020
45
39
6
5
40
34
2020
47
42
5
5
41
36
2020
50
45
5
5
45
40
2020
55
49
6
5
49
43
2020
56
51
6
6
51
46
2020
59
54
5
5
54
49
2020
64
57
6
5
59
52
2020
70
61
6
5
64
56
2020
76
69
6
6
70
63
2020
86
75
8
6
79
69
2020
97
87
8
7
89
80
2020
111
100
9
8
102
92
2020
131
119
12
11
120
109
2020
166
152
15
14
152
139
2020
29
31
5
6
24
26
2020
31
30
6
5
25
25
2020
31
31
5
6
26
26
2020
33
31
6
5
28
25
2020
34
32
5
6
29
27
2020
36
33
6
6
30
27
2020
37
34
6
6
32
29
2020
37
34
5
5
31
28
2020
38
36
5
6
33
31
2020
38
35
5
5
33
30
2020
40
37
5
6
35
32
2020
41
37
5
5
36
31
2020
43
39
5
6
37
33
2020
44
39
6
6
39
34
2020
44
39
5
5
39
34
2020
45
41
5
6
40
35
2020
47
41
6
6
42
36
2020
44
42
5
6
39
37
2020
45
42
5
6
40
37
2020
50
42
6
5
45
37
2020
50
43
5
5
45
38
2020
51
44
5
5
46
39
2020
54
46
6
6
48
41
2020
55
47
6
5
50
42
2020
57
49
6
6
51
44
2020
59
51
6
6
53
46
2020
59
52
5
6
54
47
2020
63
54
6
5
58
48
2020
65
56
6
6
59
51
2020
69
58
6
5
63
53
2020
33
32
5
5
28
27
2020
33
33
5
6
28
28
2020
35
35
6
6
30
29
2020
35
35
5
6
30
30
2020
36
34
6
5
31
29
2020
36
36
5
6
31
31
2020
38
37
6
6
33
32
2020
38
37
6
6
33
32
2020
38
37
5
5
33
32
2020
39
39
5
6
34
34
2020
41
40
5
6
36
35
2020
43
40
6
6
38
35
2020
44
41
6
6
39
36
2020
44
42
6
5
39
37
2020
45
44
6
6
40
39
2020
47
46
5
6
42
41
2020
49
46
6
5
44
41
2020
50
48
6
6
45
43
2020
52
51
5
6
47
46
2020
55
52
6
5
50
47
2020
56
53
6
5
51
48
2020
58
57
5
6
53
52
2020
61
59
6
6
56
53
2020
63
59
6
5
58
54
2020
68
65
6
6
62
60
2020
72
71
7
7
66
64
2020
81
77
8
7
74
71
2020
87
81
8
7
80
74
2020
92
85
8
7
84
78
2020
105
99
10
8
96
91
2020
34
34
5
6
29
29
2020
35
34
5
5
29
28
2020
37
35
6
5
32
29
2020
38
36
6
6
33
31
2020
38
37
6
6
33
31
2020
39
38
5
6
33
32
2020
39
38
5
6
34
33
2020
40
38
6
5
35
32
2020
42
39
6
6
37
34
2020
42
39
5
5
37
34
2020
End of preview. Expand in Dataset Viewer.

Dataset Card for GTSRB

Dataset Summary

The German Traffic Sign Benchmark is a multi-class, single-image classification challenge held at the International Joint Conference on Neural Networks (IJCNN) 2011. We cordially invite researchers from relevant fields to participate: The competition is designed to allow for participation without special domain knowledge. Our benchmark has the following properties:

  • Single-image, multi-class classification problem
  • More than 40 classes
  • More than 50,000 images in total
  • Large, lifelike database

Supported Tasks and Leaderboards

Kaggle
Original

Dataset Structure

Data Instances 

{
  "Width": 31,
  "Height": 31,
  "Roi.X1": 6,
  "Roi.Y1": 6,
  "Roi.X2": 26,
  "Roi.Y2": 26,
  "ClassId": 20,
  "Path": "Train/20/00020_00004_00002.png",
}

Data Fields

  • Width: width of image
  • Height: Height of image
  • Roi.X1: Upper left X coordinate
  • Roi.Y1: Upper left Y coordinate
  • Roi.X2: Lower right t X coordinate
  • Roi.Y2: Lower right Y coordinate
  • ClassId: Class of image
  • Path: Path of image

Data Splits

Categories: 42 Train: 39209 Test: 12630

Dataset Creation

Curation Rationale

Recognition of traffic signs is a challenging real-world problem of high industrial relevance. Although commercial systems have reached the market and several studies on this topic have been published, systematic unbiased comparisons of different approaches are missing and comprehensive benchmark datasets are not freely available.

Traffic sign recognition is a multi-class classification problem with unbalanced class frequencies. Traffic signs can provide a wide range of variations between classes in terms of color, shape, and the presence of pictograms or text. However, there exist subsets of classes (e. g., speed limit signs) that are very similar to each other.

The classifier has to cope with large variations in visual appearances due to illumination changes, partial occlusions, rotations, weather conditions, etc.

Humans are capable of recognizing the large variety of existing road signs with close to 100% correctness. This does not only apply to real-world driving, which provides both context and multiple views of a single traffic sign, but also to the recognition from single images.

Downloads last month
112
Edit dataset card
Homepage:
sciencedirect.com
Repository:
github.com
Paper:
Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognition
Leaderboard:
benchmark.ini.rub.de
Point of Contact:
bhorsey16@gmail.com
Size of downloaded dataset files:
312 MB
Size of the auto-converted Parquet files:
312 MB
Number of rows:
51,839

Models trained or fine-tuned on bazyl/GTSRB