SUN_Colonoscopy_Video.md

SUN Colonoscopy Video

Dataset Information

The SUN Colonoscopy Video dataset is a collection of endoscopic video frames for polyp detection, developed in collaboration between Nagoya University and Showa University Northern Yokohama Hospital. The dataset is designed to evaluate the effectiveness of automated colon polyp detection technologies. It includes 49,136 polyp frames from 100 different polyps, all of which have been meticulously annotated with bounding boxes. In addition, the database contains 109,554 frames of non-polyp scenes.

Automating polyp detection to alleviate the workload of endoscopists and improve the accuracy and efficiency of detection is crucial for the early diagnosis and treatment of colon cancer. The SUN Colonoscopy Video dataset advances the development of related automated detection technologies by providing a large number of real-world images with expert-level annotations, making these technologies more feasible and effective in real-world applications.

Dataset Meta Information

Dimensions	Modality	Task Type	Anatomical Structures	Number of Categories	Data Volume	File Format
2D	Endoscopy	Polyp Detection	colon	1	49,136 polyp frames, 109,554 non-polyp frames	jpg

Label Information Statistics

Polyp frame label statistics:

ID	Detection Count	Type	Size	Location	Pathological Type	Ratio (%)
1	527	Is	6mm	Cecum	Low-grade adenoma	1.0725338
2	1,313	Is	18mm	Rectum	High-grade adenoma	2.67217519
3	292	IIa	3mm	Ascending colon	Low-grade adenoma	0.59426897
4	80	Is	4mm	Sigmoid colon	Low-grade adenoma	0.16281342
5	930	IIa	3mm	Transverse colon	Low-grade adenoma	1.89270596
6	491	IIa	3mm	Sigmoid colon	Low-grade adenoma	0.99926734
7	315	IIa	6mm	Descending colon	Low-grade adenoma	0.64107782
8	256	Isp	12mm	Sigmoid colon	Low-grade adenoma	0.52100293
9	136	Is	4mm	Sigmoid colon	Low-grade adenoma	0.27678281
10	436	IIa	3mm	Transverse colon	Low-grade adenoma	0.88733312
11	113	IIa	5mm	Descending colon	Low-grade adenoma	0.22997395
12	538	Is	5mm	Rectum	Low-grade adenoma	1.09492022
13	479	Is	5mm	Transverse colon	Low-grade adenoma	0.97484533
14	1,183	IIa	3mm	Sigmoid colon	Low-grade adenoma	2.40760339
15	487	Is	5mm	Transverse colon	Low-grade adenoma	0.99112667
16	199	Is	4mm	Transverse colon	Low-grade adenoma	0.40499837
17	304	Is	4mm	Sigmoid colon	Low-grade adenoma	0.61869098
18	243	Is	2mm	Sigmoid colon	Hyperplastic polyp	0.49454575
19	96	IIa	3mm	Transverse colon	Low-grade adenoma	0.1953761
20	3,159	IIa	3mm	Ascending colon	Low-grade adenoma	6.42909476
21	100	IIa	3mm	Sigmoid colon	Low-grade adenoma	0.20351677
22	314	IIa	2mm	Ascending colon	Low-grade adenoma	0.63904266
23	182	Ip	12mm	Ascending colon	Low-grade adenoma	0.37040052
24	973	Ip	15mm-	Sigmoid colon	Low-grade adenoma	1.98021817
25	338	Is	7mm	Sigmoid colon	Low-grade adenoma	0.68788668
26	370	Is	5mm	Descending colon	Low-grade adenoma	0.75301205
27	249	Is	5mm	Ascending colon	Hyperplastic polyp	0.50675676
28	195	Is	2mm	Transverse colon	Low-grade adenoma	0.3968577
29	377	Isp	13mm	Sigmoid colon	Low-grade adenoma	0.76725822
30	224	IIa	4mm	Sigmoid colon	Low-grade adenoma	0.45587756
31	183	Ip	12mm	Descending colon	Low-grade adenoma	0.37243569
32	981	Ip	15mm-	Ascending colon	Traditional serrated adenoma	1.99649951
33	594	Is	5mm	Sigmoid colon	Low-grade adenoma	1.20888961
34	245	Is	3mm	Ascending colon	Low-grade adenoma	0.49861609
35	1,212	Ip	15mm-	Sigmoid colon	High-grade adenoma	2.46662325
36	815	IIa	7mm	Sigmoid colon	Low-grade adenoma	1.65866167
37	448	Is	7mm	Transverse colon	Low-grade adenoma	0.91175513
38	509	Is	5mm	Ascending colon	Low-grade adenoma	1.03590036
39	713	IIa	13mm	Ascending colon	Low-grade adenoma	1.45107457
40	159	IIa	5mm	Transverse colon	Low-grade adenoma	0.32359166
41	108	IIa	3mm	Rectum	Low-grade adenoma	0.21979811
42	268	Is	7mm	Transverse colon	Low-grade adenoma	0.54542494
43	260	Isp	10mm	Ascending colon	Low-grade adenoma	0.5291436
44	745	IIa	5mm	Sigmoid colon	Low-grade adenoma	1.51619993
45	383	Is	3mm	Ascending colon	Low-grade adenoma	0.77946923
46	170	IIa	2mm	Transverse colon	Hyperplastic polyp	0.34597851
47	705	Is	5mm	Transverse colon	Low-grade adenoma	1.43479323
48	176	Is	3mm	Transverse colon	Low-grade adenoma	0.35818951
49	181	IIa	3mm	Transverse colon	Low-grade adenoma	0.36836535
50	740	Ip	10mm	Sigmoid colon	Low-grade adenoma	1.5060241
51	1,737	IIa (LST-NG)	15mm	Cecum	Low-grade adenoma	3.53508629
52	207	IIa	6mm	Sigmoid colon	Low-grade adenoma	0.42127971
53	245	Is	4mm	Rectum	Hyperplastic polyp	0.49861609
54	345	Is	4mm	Sigmoid colon	Low-grade adenoma	0.70213286
55	700	Is	3mm	Ascending colon	Low-grade adenoma	1.42461739
56	248	Is	4mm	Sigmoid colon	Hyperplastic polyp	0.50472159
57	326	Is	5mm	Transverse colon	Low-grade adenoma	0.66346467
58	267	IIa	6mm	Transverse colon	Sessile serrated lesion	0.54338978
59	646	Isp	8mm	Sigmoid colon	Traditional serrated adenoma	1.31471833
60	146	IIa	8mm	Transverse colon	Low-grade adenoma	0.29713448
61	679	Isp	6mm	Ascending colon	Low-grade adenoma	1.38187887
62	351	Is	7mm	Ascending colon	Low-grade adenoma	0.71434386
63	632	Is	7mm	Rectum	Invasive cancer (T1b)	1.28622599
64	81	IIa	3mm	Sigmoid colon	Low-grade adenoma	0.16484858
65	222	IIa	3mm	Cecum	Low-grade adenoma	0.45180723
66	1,685	Is	6mm	Sigmoid colon	Low-grade adenoma	3.42925757
67	191	IIa	5mm	Transverse colon	Low-grade adenoma	0.38871703
68	1,319	Is	15mm	Rectum	High-grade adenoma	2.68438619
69	130	IIa	3mm	Descending colon	Low-grade adenoma	0.2645718
70	264	Ip	15mm	Sigmoid colon	Low-grade adenoma	0.53728427
71	1,021	Is	4mm	Ascending colon	Low-grade adenoma	2.07790622
72	774	Is	5mm	Ascending colon	Low-grade adenoma	1.5752198
73	1,285	Is	3mm	Cecum	Low-grade adenoma	2.61519049
74	276	Isp	5mm	Sigmoid colon	Low-grade adenoma	0.56170628
75	343	Is	3mm	Transverse colon	Low-grade adenoma	0.69806252
76	343	Is	3mm	Cecum	Low-grade adenoma	0.69806252
77	215	Is	4mm	Ascending colon	Low-grade adenoma	0.43756106
78	267	Isp	12mm	Sigmoid colon	High-grade adenoma	0.54338978
79	76	Is	4mm	Descending colon	Low-grade adenoma	0.15467275
80	1,192	Is	10mm	Sigmoid colon	Low-grade adenoma	2.4259199
81	427	Is	6mm	Sigmoid colon	Low-grade adenoma	0.86901661
82	111	IIa	3mm	Sigmoid colon	Sessile serrated lesion	0.22590361
83	795	Isp	13mm	Rectum	Low-grade adenoma	1.61795832
84	218	Is	5mm	Descending colon	Low-grade adenoma	0.4436656
85	1,393	IIa	8mm	Ascending colon	Low-grade adenoma	2.8349886
86	257	IIa	4mm	Sigmoid colon	Low-grade adenoma	0.5230381
87	454	Is	3mm	Cecum	Low-grade adenoma	0.92396613
88	249	Is	4mm	Ascending colon	Low-grade adenoma	0.50675676
89	149	Ip	5mm	Descending colon	Low-grade adenoma	0.3032399
90	479	Is	10mm	Ascending colon	Sessile serrated lesion	0.97484533
91	1,061	IIa	13mm	Ascending colon	Low-grade adenoma	2.15931293
92	391	Is	7mm	Descending colon	Low-grade adenoma	0.79575057
93	452	Is	7mm	Descending colon	Low-grade adenoma	0.9198958
94	136	Is	6mm	Sigmoid colon	Low-grade adenoma	0.27678281
95	606	Isp	8mm	Sigmoid colon	Low-grade adenoma	1.23331162
96	301	Is	5mm	Sigmoid colon	Hyperplastic polyp	0.61258548
97	431	IIa	15mm-	Cecum	Sessile serrated lesion	0.87715728
98	170	IIa	4mm	Transverse colon	Low-grade adenoma	0.34597851
99	161	Is	5mm	Sigmoid colon	Low-grade adenoma	0.327662
100	188	IIa	3mm	Rectum	Hyperplastic polyp	0.38261153

Non-polyp frame label statistics:

ID	Detection Count	Detection Time (s)	Ratio (%)
1	9,961	332	9.09232
2	10,073	335.8	9.194552
3	7,152	238.4	6.528287
4	14,635	487.8	13.35871
5	7,916	263.9	7.22566
6	17,046	511.4	15.55945
7	5,636	169.1	5.144495
8	2,568	85.6	2.34405
9	9,522	317.4	8.691604
10	7,086	236.2	6.468043
11	4,832	161.1	4.41061
12	6,799	226.6	6.206072
13	6,328	210.9	5.776147

Visualization

File Structure

As the official sources have not provided the specific structure of the files, and since SUN-SEG has provided data preparation for this dataset, we will use the data preparation provided by SUN-SEG as the standard for reference.

The processed data structure is as follows:

├──data
    ├──SUN-SEG
        ├──TrainDataset
            ├──Frame  # The images from the SUN dataset
                ├──case1_1
                    ├──image_name_00001.jpg
                    |...
                ├──case1_3
                |...
            ├──GT  # Object-level segmentation mask
                ├──case1_1
                    ├──image_name_00001.png
                    |...
                ├──case1_3
                |...
            ├──Edge  # Weak label with edge
                |...
            ├──Scribble  # Weak label with scribble
                |...
            ├──Polygon  # Weak label with Polygon
                |...
            ├──Classification  # Category classification annotation
                ├──classification.txt
            ├──Detection  # Bounding box
                ├──bbox_annotation.json
        ├──TestEasyDataset
            ├──Seen
                ├──Frame
                    ├──case2_3
                    |...
                ├──GT
                    ├──case2_3
                    |...
                |...
            ├──Unseen
                ├──Frame
                    ├──case3_1
                    |...
                ├──GT
                    ├──case3_1
                    |...
                |...
        ├──TestHardDataset
            ├──Seen
                ├──Frame
                    ├──case1_2
                    |...
                ├──GT
                    ├──case1_2
                    |...
                |...
            ├──Unseen
                ├──Frame
                    ├──case10_1
                    |...
                ├──GT
                    ├──case10_1
                    |...
                |...

Authors and Institutions

Masashi Misawa (Digestive Disease Center, Showa University Northern Yokohama Hospital)

Shin-Ei Kudo (Digestive Disease Center, Showa University Northern Yokohama Hospital)

Yuichi Mori (Digestive Disease Center, Showa University Northern Yokohama Hospital)

Kinichi Hotta (Endoscopy Division, Shizuoka Cancer Center)

Kazuo Ohtsuka (Endoscopy Department, Tokyo Medical and Dental University)

Takahisa Matsuda (Cancer Screening Center, National Cancer Center Hospital; Division of Screening Technology, Center for Public Health Sciences, National Cancer Center; Endoscopy Division, National Cancer Center Hospital)

Shoichi Saito (Gastroenterology Department, Cancer Institute Hospital)

Toyoki Kudo (Digestive Disease Center, Showa University Northern Yokohama Hospital)

Toshiyuki Baba (Digestive Disease Center, Showa University Northern Yokohama Hospital)

Fumio Ishida (Digestive Disease Center, Showa University Northern Yokohama Hospital)

Hayato Itoh (Graduate School of Informatics, Nagoya University)

Masahiro Oda (Graduate School of Informatics, Nagoya University)

Kensaku Mori (Graduate School of Informatics, Nagoya University)

Source Information

Official Website: http://amed8k.sundatabase.org/

Download Link: http://amed8k.sundatabase.org/

Article Address: https://www.giejournal.org/article/S0016-5107(20)34655-1/fulltext

Publication Date: 2020-7

Citation

@article{misawa2021development,
  title={Development of a computer-aided detection system for colonoscopy and a publicly accessible large colonoscopy video database (with video)},
  author={Misawa, Masashi and Kudo, Shin-ei and Mori, Yuichi and Hotta, Kinichi and Ohtsuka, Kazuo and Matsuda, Takahisa and Saito, Shoichi and Kudo, Toyoki and Baba, Toshiyuki and Ishida, Fumio and others},
  journal={Gastrointestinal endoscopy},
  volume={93},
  number={4},
  pages={960--967},
  year={2021},
  publisher={Elsevier}
}

Original introduction article is here.