ImageClassifier

Aiwozo.AI.Core.Activities.ImageClassifier

• ImageClassifier activity is used for classifying the image of different kinds by extracting the features from an image.
• Models are trained with the properly classified image dataset to predict the future coming images according to the database.
• Following the steps for configuring the ImageClassifier:
  • In the Data section, images need to be imported to the classes which required to be trained.
  • In the Train section, Model must be selected (from Auto, Logistic Regression, Support Vector Machine, Random Forest & Neural Network Model) for which the classes need to be trained by providing the Image width, Image Height, Test Size.
  • Auto model is selected by default. For other Model except Auto, some extra parameters need to be defined which are as follows:

1. Logistic Regression –
   • Penalty –
   • Dual –
   • Tolerance –
   • Penalty Strength –
2. Support Vector Machine –
   • Penalty Strength –
   • Kernel –
   • Polynomial Degree –
   • Tolerance –
3. Random Forest –
   • N Estimators –
   • Criterion –
   • Max Depth –
   • N Jobs –
4. Neural Network Model –
   • Ephocs –
   • Activation –
   • Optimizer –
   • Loss –

• In the Test section, images can be added for the testing. Testing result is the confidence of the image for each class the model is trained on.

 Properties

      Misc

• DisplayName (String) – To set a meaningful name according to the user.
• Output (OutArgument<String>) –
• path (String) –
• Result (OutArgument<String>) –