SchizoGoogLeNet: The GoogLeNet-Based Deep Feature Extraction Design for Automatic Detection of Schizophrenia

Table 1

Model architecture of GoogLeNet used in this study.

Layer	Patch size/stride	Depth	#1 × 1	#3 × 3 reduce	#3 × 3	#5 × 5 reduce	#5 × 5	Pool proj	Output size
Conv1	7 × 7/2	1							112 × 112 × 64
Max pool1	3 × 3/2	0							56 × 56 × 64
Conv2	3 × 3/1	2		64	192				56 × 56 × 192
Max pool2	3 × 3/2	0							28 × 28 × 192
Inception-3a		2	64	96	128	16	32	32	28 × 28 × 256
Inception-3b		2	128	128	192	32	96	64	28 × 28 × 480
Max pool3	3 × 3/2	0							14 × 14 × 480
Inception-4a		2	192	96	208	16	48	64	14 × 14 × 512
Inception-4b		2	160	112	224	24	64	64	14 × 14 × 512
Inception-4c		2	128	128	256	24	64	64	14 × 14 × 512
Inception-4d		2	112	144	288	32	64	64	14 × 14 × 528
Inception-4e		2	256	160	320	32	128	128	14 × 14 × 832
Max pool4	3 × 3/2	0							7 × 7 × 832
Inception-5a		2	256	160	320	32	128	128	7 × 7 × 832
Inception-5b		2	384	192	384	48	128	128	7 × 7 × 1024
Average pool5	7 × 7/1	0							1 × 1 × 1024
Dropout (40%)		0							1 × 1 × 1024
Fc		1							1 × 1 × 2
Softmax		0							1 × 1 × 2
Classification output									1 × 1 × 2