Using parallel pre-trained types of DCNN model to predict breast cancer with color normalization

BMC Research Notes

Table 1 Changing the statistical values of the suggested model (SM) in comparison with the Densenet model (DM) and the mean average of both models together (AM)

Threshold (%)	Mis class benign			Mis class malignant			Mis class			Accuracy
Threshold (%)	SM	DM	AM	SM	DM	AM	SM	DM	AM	SM (%)	DM (%)	AM (%)
50	26	34	26	14	19	15	40	53	41	97	97	97
55	23	29	23	14	21	17	37	50	40	98	97	97
60	20	25	19	14	21	18	34	46	37	98	97	98
65	19	25	15	16	22	23	35	47	38	98	97	98
70	18	24	13	18	24	24	36	48	37	98	97	98
75	19	19	13	17	27	26	36	46	39	98	97	98
80	15	19	11	23	29	32	38	48	43	98	97	97
85	15	18	10	30	34	39	45	52	49	97	97	97
90	11	15	8	33	39	41	44	54	49	97	97	97
95	11	10	5	54	49	55	65	59	60	96	96	96
97	8	7	2	57	60	65	65	67	67	96	96	96

ISSN: 1756-0500