Tumor Segmentation in Contrast-Enhanced Magnetic Resonance Imaging for Nasopharyngeal Carcinoma: Deep Learning with Convolutional Neural Network

Table 1

Detailed parameters of the CNN network.

Block	Layer	Kernel size	Stride	Pad	Output size
Input Data	-	-	-	-	1441441
Conv-Relu 1	Conv1	5	1	2	14414480
	Relu1	-	-	-	14414480
Pool-Conv-Relu 1	Pool1	2	2	-	727280
	Conv2	5	1	2	7272120
	Relu2	-	-	-	7272120
Pool-Conv-Relu 2	Pool2	2	2	-	3636120
	Conv3	9	1	4	3636500
	Relu3	-	-	-	3636500
Conv-Relu 2	Conv4	1	1	0	3636300
	Relu4	-	-	-	3636300
Conv-Relu 3	Conv5	3	1	1	3636100
	Relu5	-	-	-	3636100
Conv-Relu 4	Conv6	3	1	1	36361
	Relu6	-	-	-	36361
Deconv1	Deconv1	4	2	1	72721
Concatenate1	Concat1	-	-	-	7272121
Conv-Relu 5	Conv7	1	1	0	72721
	Relu7	-	-	-	72721
Deconv2	Deconv2	4	2	1	1441441
Concatenate2	Concat2	-	-	-	14414481
Conv6	Conv8	1	1	0	1441441
Output	Flatten1	-	-	-	1441441