Comment on page

Choose Folder 2D

Choose Folder 2D Node Documentation
The Choose Folder 2D node in AugeLab Studio allows you to choose a structured folder containing images for a 2D CNN model. It is specifically designed for image classification tasks where the images are organized into subfolders representing different classes.
Node Overview
The Choose Folder 2D node allows you to select a structured folder containing images and configure various parameters for data loading. It provides the following outputs:
PNNModel: The configuration object that contains the information about the chosen folder, class labels, and other parameters for data loading.
Node Properties
Node Title: Choose Folder 2D
Node ID: OP_NODE_AI_2D_CHOOSE_FOLDER
Inputs
The Choose Folder 2D node does not have any input sockets.
Outputs
The Choose Folder 2D node has one output socket:
1.PNNModel: The configuration object containing the information about the chosen folder, class labels, and other parameters for data loading.
Node Interaction
1.Drag and drop the Choose Folder 2D node from the node library onto the canvas in AugeLab Studio.
2.Click on the "Choose Folder" button to select the folder containing the image data.
3.Navigate to the desired folder and click "OK" to select it.
4.The selected folder will be displayed in the node.
5.Configure the desired options for channel size and data augmentation.
6.Run the pipeline.
7.The output socket of the Choose Folder 2D node will provide the configuration object (PNNModel) containing the information about the chosen folder, class labels, and other parameters for data loading.
Implementation Details
The Choose Folder 2D node allows you to select a structured folder containing images for a 2D CNN model. It provides options to choose the channel size (grayscale or RGB) and enable data augmentation.
When the "Choose Folder" button is clicked, a file dialog will open, allowing you to navigate to the desired folder. The selected folder should have the following structure:
- Folder
  - train
    - Class 1
      - Image 1
      - Image 2
      ...
    - Class 2
      - Image 1
      - Image 2
      ...
    ...
  - validation
    - Class 1
      - Image 1
      - Image 2
      ...
    - Class 2
      - Image 1
      - Image 2
      ...
    ...
  - test
    - Image 1
    - Image 2
    ...
The Choose Folder 2D node reads the folder structure and extracts the class labels from the subfolder names in the "train" directory. It creates a configuration object (PNNModel) that stores the folder path, class labels, color type (grayscale or RGB), and data augmentation settings.
The output configuration object (PNNModel) can be used as input to other nodes in the pipeline for training or evaluation of a 2D CNN model.
Usage
1.Drag and drop the Choose Folder 2D node from the node library onto the canvas in AugeLab Studio.
2.Click on the "Choose Folder" button to select the folder containing the image data.
3.Navigate to the desired folder and click "OK" to select it.
4.The selected folder will be displayed in the node.
5.Configure the desired options for channel size and data augmentation.
6.Run the pipeline.
7.The output socket of the Choose Folder 2D node will provide the configuration object (PNNModel) containing the information about the chosen folder, class labels, and other parameters for data loading.
Notes
The Choose Folder 2D node is designed for image classification tasks where the images are organized into subfolders representing different classes.
The selected folder should have the appropriate structure with "train", "validation", and "test" subfolders.
Ensure that the folder structure and image files are correctly set up before running the pipeline.
The Choose Folder 2D node reads the class labels from the subfolder names in the "train" directory.
The color type can be set to either grayscale or RGB, depending on the input images.
The Choose Folder 2D node supports data augmentation by enabling the "Data Augmentation" checkbox.

Function Blocks - AI Model Elements - Previous

Batch Normalization

Next - Function Blocks - AI Model Elements

Compile Model

Last modified 4mo ago