Skip to main content
. 2023 Jul 6;23(13):6195. doi: 10.3390/s23136195
Algorithm 2 Event-based dataset testing on NN architecture
  • Input: 

    raw events from dataset

  • Output: 

    FZE, FPE

  •      

    INITIALIZATION:

  •   1:

    load trained weight of network model

  •   2:

    load raw events

  •   3:

    initialize a matrix with zeros (image)

  •      

    DATA PREPROCESSING:

  •   4:

    resize raw events to match input

  •      

    LOOP PROCESS

  •   5:

    for i=1 to events count do

  •   6:

         update the image with event i

  •   7:

         if (i mod #eventPerGroup) = 0 then

  •   8:

           test the image in the neural network

  •   9:

           if class_code is correct & FZE flag = 0 then

  • 10:

              FZE equals i

  • 11:

              FZE flag equals 1

  • 12:

           if class_code is correct & probability >= 0.95 & FPE flag = 0 then

  • 13:

              FPE equals i

  • 14:

              FPE flag equals 1

  • 15:

    calculate the difference between FZE and FPE