|
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 to do
-
6:
update the image with event i
-
7:
if = 0 then
-
8:
test the image in the neural network
-
9:
if & = 0 then
-
10:
FZE equals i
-
11:
FZE flag equals 1
-
12:
if & & = 0 then
-
13:
FPE equals i
-
14:
FPE flag equals 1
-
15:
calculate the difference between FZE and FPE
|