Datasets:

dronefreak
/

CeyMo

file_name stringlengths 5 8	objects dict
1.jpg	{ "bbox": [ [ 851, 880, 103, 138 ], [ 1640, 839, 278, 39 ] ], "categories": [ 4, 7 ] }
10.jpg	{ "bbox": [ [ 207, 794, 1177, 113 ] ], "categories": [ 5 ] }
100.jpg	{ "bbox": [ [ 344, 890, 152, 26 ], [ 822, 928, 223, 42 ] ], "categories": [ 2, 2 ] }
100a.jpg	{ "bbox": [ [ 476, 861, 136, 19 ], [ 854, 883, 168, 23 ] ], "categories": [ 2, 2 ] }
100b.jpg	{ "bbox": [ [ 225, 916, 165, 35 ], [ 786, 976, 283, 67 ] ], "categories": [ 2, 2 ] }
101.jpg	{ "bbox": [ [ 475, 796, 1249, 80 ] ], "categories": [ 5 ] }
101a.jpg	{ "bbox": [ [ 652, 778, 678, 39 ] ], "categories": [ 5 ] }
101b.jpg	{ "bbox": [ [ 299, 833, 1359, 110 ] ], "categories": [ 5 ] }
102.jpg	{ "bbox": [ [ 363, 903, 207, 41 ], [ 996, 942, 260, 60 ] ], "categories": [ 2, 2 ] }
102a.jpg	{ "bbox": [ [ 528, 854, 160, 24 ], [ 982, 869, 179, 31 ] ], "categories": [ 2, 2 ] }
103.jpg	{ "bbox": [ [ 261, 919, 187, 38 ], [ 849, 965, 291, 60 ] ], "categories": [ 2, 2 ] }
103a.jpg	{ "bbox": [ [ 485, 867, 144, 20 ], [ 890, 880, 183, 27 ] ], "categories": [ 2, 2 ] }
103b.jpg	{ "bbox": [ [ 351, 897, 175, 32 ], [ 867, 926, 246, 43 ] ], "categories": [ 2, 2 ] }
104a.jpg	{ "bbox": [ [ 908, 883, 174, 25 ] ], "categories": [ 2 ] }
104b.jpg	{ "bbox": [ [ 864, 971, 279, 64 ] ], "categories": [ 2 ] }
105.jpg	{ "bbox": [ [ 406, 792, 911, 75 ] ], "categories": [ 5 ] }
106.jpg	{ "bbox": [ [ 295, 879, 163, 25 ], [ 758, 913, 222, 40 ] ], "categories": [ 2, 2 ] }
106a.jpg	{ "bbox": [ [ 393, 857, 148, 20 ], [ 792, 880, 183, 25 ] ], "categories": [ 2, 2 ] }
106b.jpg	{ "bbox": [ [ 163, 905, 168, 38 ], [ 702, 969, 277, 58 ] ], "categories": [ 2, 2 ] }
107.jpg	{ "bbox": [ [ 350, 877, 164, 26 ], [ 845, 904, 183, 40 ] ], "categories": [ 2, 2 ] }
107a.jpg	{ "bbox": [ [ 451, 850, 150, 22 ], [ 863, 868, 156, 28 ] ], "categories": [ 2, 2 ] }
108.jpg	{ "bbox": [ [ 203, 798, 1615, 134 ] ], "categories": [ 5 ] }
108a.jpg	{ "bbox": [ [ 464, 788, 1089, 47 ] ], "categories": [ 5 ] }
109.jpg	{ "bbox": [ [ 399, 882, 183, 32 ], [ 870, 899, 215, 40 ] ], "categories": [ 2, 2 ] }
109a.jpg	{ "bbox": [ [ 517, 844, 151, 19 ], [ 895, 850, 168, 24 ] ], "categories": [ 2, 2 ] }
109b.jpg	{ "bbox": [ [ 264, 914, 196, 44 ], [ 843, 952, 274, 59 ] ], "categories": [ 2, 2 ] }
10a.jpg	{ "bbox": [ [ 483, 774, 741, 51 ] ], "categories": [ 5 ] }
11.jpg	{ "bbox": [ [ 393, 793, 1524, 105 ] ], "categories": [ 5 ] }
110.jpg	{ "bbox": [ [ 238, 802, 1681, 121 ] ], "categories": [ 5 ] }
110a.jpg	{ "bbox": [ [ 463, 790, 1249, 58 ] ], "categories": [ 5 ] }
110b.jpg	{ "bbox": [ [ 8, 808, 1911, 225 ] ], "categories": [ 5 ] }
111.jpg	{ "bbox": [ [ 197, 805, 1722, 167 ] ], "categories": [ 5 ] }
112.jpg	{ "bbox": [ [ 389, 873, 178, 31 ], [ 850, 910, 220, 36 ] ], "categories": [ 2, 2 ] }
112a.jpg	{ "bbox": [ [ 504, 847, 145, 22 ], [ 867, 868, 181, 24 ] ], "categories": [ 2, 2 ] }
112b.jpg	{ "bbox": [ [ 246, 901, 193, 44 ], [ 812, 964, 296, 63 ] ], "categories": [ 2, 2 ] }
113.jpg	{ "bbox": [ [ 65, 807, 1661, 208 ] ], "categories": [ 5 ] }
114.jpg	{ "bbox": [ [ 853, 846, 87, 94 ] ], "categories": [ 7 ] }
114a.jpg	{ "bbox": [ [ 887, 818, 56, 55 ] ], "categories": [ 7 ] }
114b.jpg	{ "bbox": [ [ 806, 873, 124, 154 ] ], "categories": [ 7 ] }
115.jpg	{ "bbox": [ [ 841, 860, 101, 103 ] ], "categories": [ 7 ] }
115a.jpg	{ "bbox": [ [ 894, 822, 62, 45 ], [ 950, 774, 25, 11 ] ], "categories": [ 7, 7 ] }
115b.jpg	{ "bbox": [ [ 867, 842, 83, 71 ], [ 943, 776, 30, 13 ] ], "categories": [ 7, 7 ] }
115c.jpg	{ "bbox": [ [ 797, 885, 143, 158 ], [ 937, 785, 34, 16 ] ], "categories": [ 7, 7 ] }
116a.jpg	{ "bbox": [ [ 909, 819, 62, 44 ], [ 950, 780, 35, 16 ] ], "categories": [ 7, 7 ] }
116b.jpg	{ "bbox": [ [ 889, 841, 78, 68 ], [ 948, 787, 37, 21 ] ], "categories": [ 7, 7 ] }
116c.jpg	{ "bbox": [ [ 834, 885, 130, 155 ], [ 941, 798, 44, 28 ] ], "categories": [ 7, 7 ] }
117.jpg	{ "bbox": [ [ 895, 858, 93, 111 ] ], "categories": [ 7 ] }
117a.jpg	{ "bbox": [ [ 922, 816, 59, 51 ] ], "categories": [ 7 ] }
117b.jpg	{ "bbox": [ [ 908, 840, 74, 79 ] ], "categories": [ 7 ] }
117c.jpg	{ "bbox": [ [ 877, 870, 116, 161 ] ], "categories": [ 7 ] }
117d.jpg	{ "bbox": [ [ 882, 891, 121, 164 ] ], "categories": [ 7 ] }
118.jpg	{ "bbox": [ [ 15, 840, 218, 32 ], [ 243, 795, 361, 54 ] ], "categories": [ 7, 0 ] }
118a.jpg	{ "bbox": [ [ 244, 819, 157, 24 ], [ 367, 785, 301, 38 ] ], "categories": [ 7, 0 ] }
119.jpg	{ "bbox": [ [ 1, 801, 657, 131 ] ], "categories": [ 3 ] }
119a.jpg	{ "bbox": [ [ 0, 818, 594, 130 ] ], "categories": [ 3 ] }
11a.jpg	{ "bbox": [ [ 570, 789, 733, 53 ] ], "categories": [ 5 ] }
11b.jpg	{ "bbox": [ [ 170, 830, 1503, 174 ] ], "categories": [ 5 ] }
12.jpg	{ "bbox": [ [ 970, 865, 143, 95 ] ], "categories": [ 6 ] }
121.jpg	{ "bbox": [ [ 272, 792, 356, 55 ], [ 78, 850, 217, 40 ] ], "categories": [ 0, 7 ] }
121a.jpg	{ "bbox": [ [ 334, 821, 128, 26 ], [ 413, 783, 269, 36 ] ], "categories": [ 7, 0 ] }
121b.jpg	{ "bbox": [ [ 5, 815, 506, 107 ] ], "categories": [ 0 ] }
122a.jpg	{ "bbox": [ [ 155, 844, 196, 35 ], [ 343, 794, 328, 49 ] ], "categories": [ 7, 0 ] }
122b.jpg	{ "bbox": [ [ 31, 807, 527, 107 ] ], "categories": [ 0 ] }
123.jpg	{ "bbox": [ [ 300, 795, 991, 78 ] ], "categories": [ 5 ] }
123a.jpg	{ "bbox": [ [ 512, 781, 665, 40 ] ], "categories": [ 5 ] }
123b.jpg	{ "bbox": [ [ 0, 814, 1500, 169 ] ], "categories": [ 5 ] }
124.jpg	{ "bbox": [ [ 125, 839, 203, 32 ], [ 300, 791, 359, 54 ] ], "categories": [ 7, 0 ] }
124a.jpg	{ "bbox": [ [ 398, 812, 112, 20 ], [ 455, 783, 268, 31 ] ], "categories": [ 7, 0 ] }
124b.jpg	{ "bbox": [ [ 42, 810, 498, 88 ] ], "categories": [ 0 ] }
125.jpg	{ "bbox": [ [ 879, 896, 217, 39 ], [ 472, 876, 169, 31 ], [ 165, 830, 120, 27 ], [ 292, 814, 141, 25 ], [ 384, 782, 296, 37 ] ], "categories": [ 2, ...
125a.jpg	{ "bbox": [ [ 900, 856, 156, 21 ], [ 576, 842, 148, 19 ], [ 301, 816, 112, 19 ], [ 404, 803, 105, 18 ], [ 456, 774, 265, 30 ] ], "categories": [ 2, ...
125b.jpg	{ "bbox": [ [ 843, 959, 294, 62 ], [ 330, 920, 205, 42 ], [ 16, 843, 113, 38 ], [ 163, 827, 181, 29 ], [ 319, 787, 323, 44 ] ], "categories": [ 2, ...
126.jpg	{ "bbox": [ [ 129, 835, 194, 27 ], [ 300, 796, 345, 45 ] ], "categories": [ 7, 0 ] }
126a.jpg	{ "bbox": [ [ 355, 809, 122, 20 ], [ 435, 782, 260, 31 ] ], "categories": [ 7, 0 ] }
127.jpg	{ "bbox": [ [ 3, 795, 681, 122 ] ], "categories": [ 3 ] }
127a.jpg	{ "bbox": [ [ 105, 790, 537, 82 ] ], "categories": [ 3 ] }
128.jpg	{ "bbox": [ [ 21, 796, 494, 90 ] ], "categories": [ 0 ] }
128a.jpg	{ "bbox": [ [ 288, 814, 142, 21 ], [ 400, 778, 282, 39 ] ], "categories": [ 7, 0 ] }
128b.jpg	{ "bbox": [ [ 17, 846, 226, 34 ], [ 232, 796, 380, 59 ] ], "categories": [ 7, 0 ] }
12a.jpg	{ "bbox": [ [ 653, 819, 124, 32 ], [ 991, 818, 91, 37 ] ], "categories": [ 7, 6 ] }
12b.jpg	{ "bbox": [ [ 623, 835, 114, 39 ], [ 978, 834, 114, 55 ] ], "categories": [ 7, 6 ] }
13.jpg	{ "bbox": [ [ 1280, 837, 110, 57 ] ], "categories": [ 6 ] }
130.jpg	{ "bbox": [ [ 0, 770, 722, 175 ] ], "categories": [ 3 ] }
130a.jpg	{ "bbox": [ [ 3, 766, 818, 119 ] ], "categories": [ 3 ] }
130b.jpg	{ "bbox": [ [ 0, 764, 784, 194 ] ], "categories": [ 3 ] }
132.jpg	{ "bbox": [ [ 0, 809, 487, 91 ] ], "categories": [ 0 ] }
132a.jpg	{ "bbox": [ [ 181, 822, 186, 33 ], [ 328, 781, 328, 44 ] ], "categories": [ 7, 0 ] }
132b.jpg	{ "bbox": [ [ 239, 795, 382, 55 ], [ 11, 843, 235, 36 ] ], "categories": [ 0, 7 ] }
134.jpg	{ "bbox": [ [ 927, 823, 70, 61 ], [ 546, 816, 156, 52 ], [ 263, 803, 209, 31 ] ], "categories": [ 7, 7, 9 ] }
134a.jpg	{ "bbox": [ [ 555, 789, 89, 16 ], [ 738, 794, 64, 21 ], [ 941, 797, 44, 21 ] ], "categories": [ 9, 7, 7 ] }
134b.jpg	{ "bbox": [ [ 411, 795, 146, 23 ], [ 650, 805, 100, 32 ], [ 934, 810, 56, 34 ] ], "categories": [ 9, 7, 7 ] }
134c.jpg	{ "bbox": [ [ 89, 825, 281, 38 ], [ 404, 848, 230, 78 ], [ 922, 859, 81, 99 ] ], "categories": [ 9, 7, 7 ] }
134d.jpg	{ "bbox": [ [ 317, 859, 287, 99 ], [ 3, 832, 316, 42 ], [ 913, 875, 100, 133 ] ], "categories": [ 7, 9, 7 ] }
135.jpg	{ "bbox": [ [ 231, 814, 226, 35 ], [ 527, 830, 165, 57 ], [ 938, 836, 70, 67 ], [ 1223, 836, 147, 58 ] ], "categories": [ 9, 7, 7, 6 ] }
135a.jpg	{ "bbox": [ [ 495, 789, 112, 18 ], [ 706, 796, 75, 24 ], [ 942, 800, 51, 26 ], [ 1130, 802, 69, 24 ] ], "categories": [ 9, 7, 7, 6 ] }
135b.jpg	{ "bbox": [ [ 366, 802, 166, 27 ], [ 626, 812, 111, 38 ], [ 942, 816, 60, 43 ], [ 1175, 818, 102, 38 ], [ 955, 767, 31, 11 ], [ 825, 766, ...
135d.jpg	{ "bbox": [ [ 1, 843, 293, 45 ], [ 272, 873, 319, 114 ], [ 929, 883, 110, 166 ], [ 1328, 878, 300, 124 ], [ 952, 782, 40, 18 ], [ 783, 781, ...
136.jpg	{ "bbox": [ [ 198, 819, 240, 37 ], [ 515, 837, 171, 63 ], [ 943, 842, 79, 77 ], [ 585, 779, 83, 16 ], [ 767, 785, 53, 19 ], [ 953, 787, ...
136a.jpg	{ "bbox": [ [ 432, 797, 141, 24 ], [ 674, 805, 90, 32 ], [ 950, 809, 54, 36 ], [ 656, 769, 63, 13 ], [ 803, 771, 47, 16 ], [ 954, 773, 3...
136b.jpg	{ "bbox": [ [ 317, 807, 193, 31 ], [ 592, 820, 135, 46 ], [ 944, 825, 68, 53 ], [ 955, 781, 36, 16 ], [ 786, 779, 48, 16 ], [ 616, 776, ...

End of preview. Expand in Data Studio

CeyMo: Road Marking Detection Dataset

CeyMo Dataset Banner

Unofficial redistribution of the CeyMo road-marking detection dataset (WACV 2022), reformatted into a standardized YOLO-compatible layout, under the original MIT license.

Disclaimer

This repository is not an official release of the CeyMo dataset.

CeyMo was created by Oshada Jayasinghe, Sahan Hemachandra, Damith Anhettigama, Shenali Kariyawasam, Ranga Rodrigo, and Peshala Jayasekara, who retain all copyright and intellectual property rights (to the extent applicable under the repository's MIT license — see License below). This repository does not claim ownership of any images, annotations, or metadata.

This repository exists for two purposes:

To reorganize the dataset into a standardized YOLO/Ultralytics-compatible directory structure that can be used directly by many modern object detection training pipelines, with a validation split (the official release has none).
To provide a more reliable download source, as the original release is hosted on Google Drive, which can be slow or rate-limited.

This redistribution is sourced directly from the official Google Drive train.zip and test.zip linked from the official repository, not from a third-party mirror.

Dataset Overview

CeyMo is a road-marking detection benchmark covering a wide variety of challenging urban, sub-urban, and rural road scenarios. It contains 2,887 images at 1920x1080 resolution with 4,706 road-marking instances across 11 classes (arrows, pedestrian crossings, bus and cycle lanes, junction boxes, diamonds, and "slow" markings). Road markings are flat, perspective-distorted ground-plane objects, unlike the upright vehicles and signs most driving benchmarks target.

The official test set is divided into six scenario categories — normal, crowded, dazzle light, night, rain, and shadow. That per-image scenario label is not carried into this repository (see below).

Changes from the Official Release

Annotation format converted. The official ground truth for detection is per-image Pascal-VOC XML (bbox_annotations/*.xml) with class codes such as SA or PC. This repository converts every box into YOLO's normalized class x_center y_center width height format (one .txt per image) and into absolute-pixel COCO [x, y, w, h] boxes in a metadata.jsonl per split. Boxes were re-expressed, not resized or altered.
Class codes renamed to readable names (e.g. SA → straight_arrow, PC → pedestrian_crossing); see the class table below.
A validation split was added. The official release has only train (2,099 images) and test (788 images). This repository keeps the official test set untouched and carves a seeded random 15% validation split (315 images, seed 42) out of the official train set, leaving 1,784 training images. The split is random at the image level; because the images come from driving footage, near-duplicate neighboring frames may fall on both sides of the train/validation boundary, so validation scores can be optimistic. Use the official test split for headline numbers.
Not included: the official polygon annotations (JSON), pixel-level segmentation masks (PNG), the per-image camera/vehicle fields, the test-set scenario category, and the unlabeled raw video clips. Only bounding boxes are redistributed.
No image pixel content was modified. All 4,706 official boxes are preserved (none dropped).

Dataset Structure

<repo>/
├── README.md
├── ceymo_banner.jpg
└── data/
    ├── data.yaml
    ├── images/
    │   ├── train/   (*.jpg + metadata.jsonl)
    │   ├── valid/   (*.jpg + metadata.jsonl)
    │   └── test/    (*.jpg + metadata.jsonl)
    └── labels/
        ├── train/   (*.txt, mirrors images)
        ├── valid/
        └── test/

where:

data/images/<split>/ contains the 1920x1080 RGB road images, plus a metadata.jsonl (file_name and objects.bbox as absolute-pixel COCO [x, y, w, h] with objects.categories) that drives the Hugging Face dataset viewer.
data/labels/<split>/ contains one YOLO-format .txt annotation file per image (class x_center y_center width height, normalized), mirroring the image layout.
data/data.yaml is the Ultralytics dataset configuration file (class names, split paths, relative to data/).
Splits: train 1,784 images / 2,963 boxes · valid 315 images / 525 boxes · test 788 images / 1,218 boxes (2,887 images / 4,706 boxes total — matches the paper's published totals).

Classes (11)

id	class name	official code	instances (all splits)
0	`bus_lane`	BL	191
1	`cycle_lane`	CL	82
2	`diamond`	DM	1,047
3	`junction_box`	JB	172
4	`left_arrow`	LA	162
5	`pedestrian_crossing`	PC	839
6	`right_arrow`	RA	352
7	`straight_arrow`	SA	1,440
8	`slow`	SL	100
9	`straight_left_arrow`	SLA	241
10	`straight_right_arrow`	SRA	80

The distribution is imbalanced: straight_arrow is about 31% of all instances while cycle_lane and straight_right_arrow are under 2% each. Boxes are large relative to typical driving-detection targets (median box about 0.6% of image area).

Dataset Sources

Original Paper

CeyMo: See More on Roads - A Novel Benchmark Dataset for Road Marking Detection

Oshada Jayasinghe, Sahan Hemachandra, Damith Anhettigama, Shenali Kariyawasam, Ranga Rodrigo, Peshala Jayasekara

Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), January 2022, pages 3104-3113.

Official Resources

Official Repository: https://github.com/oshadajay/CeyMo
Paper (CVF Open Access): https://openaccess.thecvf.com/content/WACV2022/papers/Jayasinghe_CeyMo_See_More_on_Roads_-_A_Novel_Benchmark_Dataset_WACV_2022_paper.pdf

Attribution

All credit for collecting and annotating this dataset belongs entirely to the original CeyMo authors: Oshada Jayasinghe, Sahan Hemachandra, Damith Anhettigama, Shenali Kariyawasam, Ranga Rodrigo, and Peshala Jayasekara.

This repository only reformats their bounding-box annotations into a YOLO-compatible layout and adds a validation split, for improved usability. It does not modify, reinterpret, or take credit for the underlying imagery or annotations.

If you use this dataset in your research, please cite the original publication below.

License

The CeyMo repository is released under the MIT License, as confirmed by GitHub's license detection on the official repository. Note that this is the license of the repository that distributes the dataset; the authors do not publish a separate data-specific license.

Accordingly:

Attribution and preservation of the copyright and license notice are required.
Commercial and non-commercial use are both permitted.
No share-alike obligation.

This repository is distributed under the same MIT license.

Citation

If you use this dataset, please cite:

@InProceedings{Jayasinghe_2022_WACV,
    author    = {Jayasinghe, Oshada and Hemachandra, Sahan and Anhettigama, Damith and Kariyawasam, Shenali and Rodrigo, Ranga and Jayasekara, Peshala},
    title     = {CeyMo: See More on Roads - A Novel Benchmark Dataset for Road Marking Detection},
    booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
    month     = {January},
    year      = {2022},
    pages     = {3104-3113}
}

Acknowledgements

We sincerely thank Oshada Jayasinghe and co-authors for creating and publicly releasing this valuable road-marking benchmark.

Downloads last month: 23

openaccess.thecvf.com