teonsen/ImageClassificationv2.0.0.1

What is this?

This package is a wrapper library of ML.NET to do image classification. You will need ImageClassification.IO packeage along with this package.

Usege

Trainer class

Trainer class is a class to train the Model (which will be two zip files) from your image data. Image data you have to prepare in advence is a foldered image data like below. Each folder names will be predicted Label names (Apple/Banana/Orange/...).

• Fruits
  • Apple
    • apple001.jpg
    • apple002.jpg
    • ...
  • Banana
  • Orange
  • ...

More images you prepare, better result you'll get.

Next, call Trainer.GenerateModel with dataDir you prepared.

// Define hyper-paramters such as Epoch and BatchSize. var hp = new HyperParameter { Epoch = 200, BatchSize = 10, LearningRate = 0.01f, eTrainerArchitecture = eTrainerArchitectures.ResnetV250, TestFraction = 0.3f, ResultsToShow = 20 }; // Train and generate the model. var results = Trainer.GenerateModel(dataDir, hp); // Save the results as HTML file.(optional) results.SaveAsHTML();

Classifier class

Classifier class is a class to get a prediction from the image file you give (=imageToClassify). Call Classifier.GetSingleImagePrediction along with the Model you've created at previus section.

string imageToClassify = @"C:\your\imageToClassify(apple_or_banana_or_orange).png"; var p = Classifier.GetSingleImagePrediction(results.Resultfiles.PipelineZip, results.Resultfiles.ModelZip, imageToClassify); Console.WriteLine($@"Predicted image label is: ""{p.PredictedLabel}"". Score:{p.HighScore}");