|
Algorithm 1 Wavelet Feature Extraction from Images |
-
1:
procedure WaveletFeatureExtraction
-
2:
Define root folders and classes for image datasets ▹ for 5 drills and 3 classes
-
3:
Create imageDatastore instances for each dataset and classes
-
4:
for each dataset combination (1 through 5) do
-
5:
Combine training sets from other four datasets
-
6:
Assign corresponding labels to the combined training set
-
7:
Set aside one dataset as the testing set
-
8:
Extract wavelet features for both training and testing sets
-
9:
Assign labels to the training and testing sets
-
10:
end for
-
11:
Save all wavelet features and labels to a file
-
12:
end procedure
-
13:
▹
-
14:
procedure getWaveletScattering2(ImageDatastore)
-
15:
Read all images from the datastore
-
16:
Define wavelet scattering for ImageSize=[224 224]
-
17:
Initialize empty array for all features
-
18:
for each image in the datastore do
-
19:
Extract wavelet features using defined wavelet scattering
-
20:
Compute the mean of the features
-
21:
Append the features to the features array
-
22:
end for
-
23:
return features array
-
24:
end procedure
|