Table 5 The description of all nasal endoscopic white light images used in this study

From: A deep learning based smartphone application for early detection of nasopharyngeal carcinoma using endoscopic images

Disease

Categories

Number of images (SZH)

Number of images (FSH)

Number of images (LZH)

Number of total images (N, %)

NPC

3526 images (N = 1171, 8.39%)

2312 images (N = 769, 5.51%)

194 images (N = 194, 1.39%)

6032 images (N = 2134, 15.29%)

AH

2468 images (N = 824, 5.06%)

2121 images (N = 706, 5.06%)

153 images (N = 153, 1.10%)

4742 images (N = 1683, 12.06%)

AR

2079 images (N = 698, 5.00%)

1288 images (N = 424, 3.04%)

112 images (N = 112, 0.80%)

3479 images (N = 1234, 8.84%)

CRP

3124 images (N = 1044, 7.48%)

797 images (N = 261, 1.87%)

130 images (N = 130, 0.93%)

4051 images (N = 1435, 10.28%)

DNS

4316 images (N = 1446, 10.36%)

2685 images (N = 892, 6.39%)

233 images (N = 223, 1.60%)

7234 images (N = 1435, 18.42%)

NOR

5931 images (N = 1981, 14.19%)

4491 images (N = 1495, 10.71%)

347 images (N = 347, 2.49%)

10,769 images (N = 3823, 27.39%)

RHI

2344 images (N = 784, 5.62%)

591 images (N = 196, 1.40%)

98 images (N = 98, 0.70%)

3033 images (N = 1078, 7.72%)

Total images

23,788 images (N = 7948, 56.94%)

14,285 images (N = 4743, 33.98%)

1267 images (N = 1267, 9.08%)

39,340 images (N = 13,958, 100%)

  1. ‘N’ means the number of patients corresponding to the images.