Table 1 Characteristics of the development set and the external test set.
From: Artificial intelligence to detect malignant eyelid tumors from photographic images
Development set | External test set | |||
|---|---|---|---|---|
Total no. of photographic images | 1,258 | 309 | ||
Total no. of qualified photographic imagesa | 1,151 | 266 | ||
Total no. of cropped images | 1, 248 | 285 | ||
No. of patients | 675 | 176 | ||
Mean age/range (years) | 49.6 (1–100) | 50.0 (2–87) | ||
No. of women (%) | 421 (62.4) | 111 (63.1) | ||
Institution | NEH | JEH and ZFPH | ||
Location of institution | East of China | East and west of China | ||
Camera model | Canon IXUS-130, Nikon COOLPIX-S7000 | FUJIFILM FinePix-F41, HUAWEI EVA-AL00 | ||
Training set | Validation set | Internal test set | ||
Benign eyelid tumorb | 668/883 (75.7) | 121/168 (72.0) | 146/197 (74.1) | 226/285 (79.3) |
Malignant eyelid tumorb | 215/883 (24.3) | 47/168 (28.0) | 51/197 (25.9) | 59/285 (20.7) |
