Classification on Digital Pathological Images of Breast Cancer Based on Deep Features of Different Levels

Table 8

Number distribution of the dataset with different magnifications.

Fold	Data type	Magnifications				Count
		40×	100×	200×	400×
1	Train	1365 (392 : 973)	1396 (410 : 986)	1394 (414 : 980)	1248 (381 : 867)	5403 (1597 : 3806)
	Test	630 (233 : 397)	685 (234 : 451)	619 (209 : 410)	572 (207 : 365)	2506 (883 : 1623)
2	Train	1405 (436 : 969)	1449 (450 : 999)	1417 (432 : 985)	1291 (425 : 866)	5562 (1743 : 3819)
	Test	590 (189 : 401)	632 (194 : 438)	596 (191 : 405)	529 (163 : 366)	2347 (737 : 1610)
3	Train	1304 (378 : 926)	1383 (389 : 994)	1343 (389 : 954)	1180 (348 : 832)	5210 (1504 : 3706)
	Test	691 (247 : 444)	698 (255 : 443)	670 (234 : 436)	640 (240 : 400)	2699 (976 : 1723)
4	Train	1439 (441 : 998)	1553 (473 : 1080)	1448 (439 : 1009)	1291 (418 : 873)	5731 (1771 : 3960)
	Test	556 (184 : 372)	528 (171 : 357)	565 (184 : 381)	529 (170 : 359)	2718 (709 : 1469)
5	Train	1422 (455 : 967)	1530 (491 : 1039)	1458 (452 : 1006)	1339 (440 : 899)	5749 (1838 : 3911)
	Test	573 (170 : 403)	551 (153 : 398)	555 (171 : 384)	481 (148 : 333)	2160 (642 : 1518)

"Train" represents the training dataset, and "Test" represents the test dataset, and the number in braces represents the number of benign cancer and malignant cancer, respectively.