Algorithm 1 Efficient Spatial Pyramid Pooling (ESPP)

Input: The input feature layer, x; The number of input channels of x, Cin; The number of output channels of x, Cout; The parameter that control the size of the model, a.

Output: The output feature layer, out.

1: Control the number of output feature channels of the convolution process by a.   2: Cout1 = Cout / a. // In general, the parameter a is set to 4 or 8, which gives excellent control over the number of parameters of the model.   3: Make an ordinary 1*1 convolution at x, get the output out1.   4: Make an ordinary 3*3 convolution at x, get the output out2.   5: The stacking of two 3*3 convolutions gives the same perceptual field as one 5*5 convolution:   6: Make an ordinary 3*3 convolution at out2, get the output out3.   7: The stacking of three 3*3 convolutions gives the same perceptual field as one 7*7 convolution:   8: Make an ordinary 3*3 convolution at out3, get the output out4.   9: Make a 3*3 atrous convolution with the atrous rate of 1 at out1, get the first branch output out5. 10: Make a 3*3 atrous convolution with the atrous rate of 3 at out2, get the second branch output out6. 11: Make a 3*3 atrous convolution with the atrous rate of 5 at out3, get the third branch output out7. 12: Make a 3*3 atrous convolution with the atrous rate of 7 at out4, get the third branch output out8. 13: Unify the four branch outputs into the same dimension: 14: Concat ([out5, out6, out7, out8], dimension), get out9. 15: Integrate feature information: 16: Make an ordinary 1*1 convolution at out1, get the shortcut. 17: net = out9*0.8 + shortcut. 18: Get the final output: 19: out = RELU (net). // add the non-linearity. 20: return out.