Currency
Crossroad
Image
Traffic information
Visul impairment
Assistive technology
Bounding Box
- Provision of service delivering traffic information for pedestrians with low vision or other disabilities.
- Development of service classifying notes and coins for people with low or impaired vision.
Datumo provides high quality data for smarter AI. As part of Datumo's Data Sponsorship Program, Datumo cooperated with wesee in building the following dataset.
Under the motto of building “technology for the marginalized 90 percent of population”, Wesee focuses on AI algorithm development based on video recognition. We are developing a traffic navigation system(Edge AI camera) that recognizes and predicts the traffic flow and open parking spaces, as part of the Smart City business for the visually impaired. Other projects for the visually impaired are ongoing and the details could be found below.
“AI quality heavily depends on the quality and quantity of datasets. We are delighted to have obtained great datasets in both terms through the project. The thought of increasing the possibility for the visually impaired to take a step forward in their lives have lifted the team’s spirit. We anticipate the datasets built from this project to be made useful not only by us, but also by others with great passion. We also thank Datumo for putting much effort in this project. ”
- 8 types of notes * 12 specific categories * 125 images per category = 12,000
- 8 types of coins * 8 specific categories * 200 images per category = 12,800
- Total of 160 categories / 24,800 images
- Labeling method: Collection of 24,800 images of appropriate notes and coins followed by bbox annotation
- 1 object per image
- Minimum resolution of 640*480 required (regardless of image orientation)
- Used 352,810 video data of pedestrians on crosswalks from AI Hub open datasets([https://aihub.or.kr/aidata/136](https://aihub.or.kr/aidata/136))
- Selected images of appropriate crosswalks directions and pedestrian lights(green/red)
- In the occasion of crosswalks occluded by other objects:
If the object is a human, bounding boxes are drawn regardless of the occlusion region
If the object is a car, bounding boxes are drawn excluding the occlusion region
Video data were collected and labeled using Cash Mission, Datumo's crowdsourcing platform.
As the host of the "AI Dataset Sponsorship Program", Datumo has also participated in the KLUE project as a sponsor.
- 8 types of notes * 12 specific categories * 125 images per category = 12,000
- 8 types of coins * 8 specific categories * 200 images per category = 12,800
- Total of 160 categories / 24,800 images
- Labeling method: Collection of 24,800 images of appropriate notes and coins followed by bbox annotation
- 1 object per image
- Minimum resolution of 640*480 required (regardless of image orientation)
Data collection and labeling were completed using the mobile version of Datumo's crowd-sourcing platform, Cash Misshion.
Collecting images of banknotes
Twenty-four photos were taken per banknote, depending on the sides, conditions, and angles.
Drawing bounding boxes around notes and coins
Sample Data
{
"version": "3.16.7",
"flags": {},
"shapes": [
{
"label": "Hundred_front",
"line_color": null,
"fill_color": null,
"points": [
[
638.1288828,
1782.7275110399999
],
[
1269.00915912,
2391.0826867200003
]
],
"shape_type": "rectangle",
"flags": {}
}
],
"lineColor": [
0,
255,
0,
128
],
"fillColor": [
255,
0,
0,
128
],
"imagePath": "100_12_1.jpg",
"imageWidth": 1908,
"imageHeight": 4032,
"imageData": null
}
100_12_1.json
{
"version": "3.16.7",
"flags": {},
"shapes": [
{
"label": "Five_Hundred_front",
"line_color": null,
"fill_color": null,
"points": [
[
621.7438183200001,
1765.46837376
],
[
1228.7220444,
2369.6449056
]
],
"shape_type": "rectangle",
"flags": {}
}
],
"lineColor": [
0,
255,
0,
128
],
"fillColor": [
255,
0,
0,
128
],
"imagePath": "500_10_1.jpg",
"imageWidth": 1908,
"imageHeight": 4032,
"imageData": null
}
cs
500_10_1.json
{
"version": "3.16.7",
"flags": {},
"shapes": [
{
"label": "Ten_Thousand_back",
"line_color": null,
"fill_color": null,
"points": [
[
614.40175456,
1400.03351488
],
[
2111.9280336,
2100.4689728
]
],
"shape_type": "rectangle",
"flags": {}
}
],
"lineColor": [
0,
255,
0,
128
],
"fillColor": [
255,
0,
0,
128
],
"imagePath": "10000_B_STUFF_180_100.jpg",
"imageWidth": 2992,
"imageHeight": 2992,
"imageData": null
}
10000_B_STUFF_180_100.json
- Used 352,810 video data of pedestrians on crosswalks from AI Hub open datasets [https://aihub.or.kr/aidata/136]
- Selected images of appropriate crosswalks directions and pedestrian lights(green/red)
- In the occasion of crosswalks occluded by other objects
: If the object is a human, bounding boxes are drawn regardless of the occlusion region
: If the object is a car, bounding boxes are drawn excluding the occlusion region
400,000 video datasets of crosswalks collected by Datumo with NIA(National Information Society Agency)
Datumo, using crowd-sourcing and similar-data filtering technology, participated in AI Hub’s open datasets project by building 400,000 video datasets of various crossroads in Korea.
Datasets of crosswalks, important pedestrian facilities necessary for those with visual or mobility impairment, have not existed in such volume before. Datumo built the datasets in hopes of upholding the rights of visually impaired pedestrians, and furthermore, supporting research and development of technology regarding transportation in general.
Data collection and labeling were completed using the mobile version of Datumo's crowd-sourcing platform, Cash Mission.
Drawing bounding boxes around pedestrian signals and crosswalks
- Crowd-workers of Cashmission(mobile ver.) collected and labeled crosswalks from given image data
Sample Data
{
"version": "4.5.6",
"flags": {},
"shapes": [
{
"label": "Zebra_Cross",
"points": [
[
1220.8955223880596,
823.1343283582089
],
[
1513.4328358208954,
864.9253731343283
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
}
],
"imagePath": "MP_KSC_000016.jpg",
"imageData": null,
"imageHeight": 1080,
"imageWidth": 1920
}MP_KSC_000016.json
{
"version": "4.5.6",
"flags": {},
"shapes": [
{
"label": "Zebra_Cross",
"points": [
[
252.7777777777778,
739.7222222222223
],
[
1920.0,
1080.0
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "R_Signal",
"points": [
[
1720.9549071618037,
486.2068965517241
],
[
1757.6322801827232,
564.8085071434278
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
}
],
"imagePath": "MP_KSC_000003.jpg",
"imageData": null,
"imageHeight": 1080,
"imageWidth": 1920
}MP_KSC_000003.json
{
"version": "4.5.6",
"flags": {},
"shapes": [
{
"label": "Zebra_Cross",
"points": [
[
748.6671993453269,
527.9496905552037
],
[
1154.3828226689714,
555.1900166336154
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "Zebra_Cross",
"points": [
[
1002.2222222222222,
471.85185185185185
],
[
1224.4444444444443,
513.3333333333333
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "R_Signal",
"points": [
[
1065.5319148936169,
324.25531914893617
],
[
1085.9574468085107,
371.9148936170213
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
}
],
"imagePath": "MP_KSC_007487.jpg",
"imageData": null,
"imageHeight": 1080,
"imageWidth": 1920
}MP_KSC_007487.json
{
"version": "4.5.6",
"flags": {},
"shapes": [
{
"label": "Zebra_Cross",
"points": [
[
0.0,
697.0
],
[
1334.6623376623377,
1019.8051948051948
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "Zebra_Cross",
"points": [
[
665.3061224489796,
542.1768707482993
],
[
976.2295081967213,
592.6229508196722
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "G_Signal",
"points": [
[
972.1088435374149,
340.13605442176873
],
[
1010.8843537414966,
414.96598639455783
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "R_Signal",
"points": [
[
885.9922178988328,
402.3346303501946
],
[
903.1128404669262,
432.6848249027238
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
}
],
"imagePath": "MP_SEL_002807.jpg",
"imageData": null,
"imageHeight": 1080,
"imageWidth": 1920
}MP_SEL_002807.json
{
"version": "4.5.7",
"flags": {},
"shapes": [
{
"label": "Zebra_Cross",
"points": [
[
0.0,
572.50933579918
],
[
1920.0,
1080.0
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "R_Signal",
"points": [
[
1839.084728953918,
142.63168214563183
],
[
1905.9308363541838,
238.51167789543234
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "Zebra_Cross",
"points": [
[
709.3457943925233,
490.6542056074766
],
[
1679.4392523364486,
553.2710280373832
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "R_Signal",
"points": [
[
761.5384615384615,
304.2735042735043
],
[
791.8803418803419,
359.4017094017094
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "R_Signal",
"points": [
[
1290.4918032786884,
320.983606557377
],
[
1320.655737704918,
358.03278688524586
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
}
],
"imagePath": "MP_SEL_009810.jpg",
"imageData": null,
"imageHeight": 1080,
"imageWidth": 1920
}MP_SEL_009810.json
{
"version": "4.5.7",
"flags": {},
"shapes": [
{
"label": "R_Signal",
"points": [
[
1508.9909847812826,
169.0344848385404
],
[
1563.7863306321428,
257.0436274784295
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "Zebra_Cross",
"points": [
[
9.349719239354818e-15,
564.7351640186691
],
[
1920.0000000000005,
1079.9999999999995
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "Zebra_Cross",
"points": [
[
640.2061855670104,
498.96907216494844
],
[
1426.8041237113403,
544.3298969072165
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "Zebra_Cross",
"points": [
[
806.1855670103093,
476.2886597938144
],
[
1261.8556701030927,
494.8453608247423
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "R_Signal",
"points": [
[
731.6239316239316,
268.8034188034188
],
[
762.8205128205128,
323.0769230769231
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
},
{
"label": "R_Signal",
"points": [
[
848.2905982905984,
325.64102564102564
],
[
876.0683760683761,
360.6837606837607
]
],
"group_id": null,
"shape_type": "rectangle",
"flags": {}
}
],
"imagePath": "MP_SEL_013519.jpg",
"imageData": null,
"imageHeight": 1080,
"imageWidth": 1920
}MP_SEL_013519.json
- Currency data
Developing AI services based on Korean notes and coins. - Crosswalks data
Developing AI services specialized in pedestrian signals or crosswalks for those with mobility impairment.
CC BY-SA
Reusers are allowed to distribute, remix, adapt, and build upon the material in any medium or format, even commercially, so long as attribution is given to the creator. If you remix, adapt, or build upon the material, you must license the modified material under identical terms.
https://creativecommons.org/licenses/by-sa/3.0/deed.en