Skip to main content

Table 1 Layer wise summary of the proposed architecture with window size 25 × 25

From: Exploring an application-oriented land-based hyperspectral target detection framework based on 3D–2D CNN and transfer learning

Layer(type)

Output Shape

Kernel_size

Stride

Parameter

Input(InputLayer)

(25,25,30,1)

0

Conv3d_1(Conv3D)

(23,23,24,8)

(7,3,3)

(1,1,1)

512

Conv3d_2(Conv3D)

(21,21,20,16)

(5,3,3)

(1,1,1)

5776

Conv3d_2(Conv3D)

(19,19,18,32)

(3,3,3)

(1,1,1)

13,856

Reshape_1(Reshape)

(19,19,576)

0

Conv2d_1(Conv2D)

(17,17,64)

(3,3)

(1,1)

331,840

Flatten_1(Flatten)

(18,496)

0

Dense_1(Dense)

(256)

4,735,232

Dropout_1(Dropout)

(256)

0

Dense_2(Dense)

(128)

32,896

Dropout_2(Dropout)

(128)

0

Dense_3(Dense)

(16)

2064