Skip to main content
. 2024 Feb 7;24(4):1092. doi: 10.3390/s24041092
Algorithm 2 CNN Feature Extraction Using ResNet-18’s 35th/68th Layer
  • 1:

    procedure CNNFeatureExtraction

  • 2:

        Initialize root folders for each dataset         ▹ for 5 drills

  • 3:

        Define categories: Green, Yellow, Red      ▹ for three subfolders

  • 4:

        Create imageDatastore instances for each dataset

  • 5:

        Combine datasets for training and testing

  • 6:

        Load pretrained ResNet-18 model

  • 7:

        Define feature extraction layer (e.g., 68th layer or 35th layer)

  • 8:

        for each combined dataset (1 through 5) do

  • 9:

            Read image

  • 10:

            Preprocess image according to CNN input requirements

  • 11:

            Extract features using the specified CNN layer

  • 12:

            Store extracted features

  • 13:

        end for

  • 14:

        Output the set of extracted features

  • 15:

    end procedure