OnnxCatalog.DnnFeaturizeImage Method

Definition

Create DnnImageFeaturizerEstimator, which applies one of the pre-trained DNN models in DnnImageModelSelector to featurize an image.

public static Microsoft.ML.Transforms.Onnx.DnnImageFeaturizerEstimator DnnFeaturizeImage (this Microsoft.ML.TransformsCatalog catalog, string outputColumnName, Func<Microsoft.ML.Transforms.Onnx.DnnImageFeaturizerInput,Microsoft.ML.Data.EstimatorChain<Microsoft.ML.Transforms.ColumnCopyingTransformer>> modelFactory, string inputColumnName = default);
static member DnnFeaturizeImage : Microsoft.ML.TransformsCatalog * string * Func<Microsoft.ML.Transforms.Onnx.DnnImageFeaturizerInput, Microsoft.ML.Data.EstimatorChain<Microsoft.ML.Transforms.ColumnCopyingTransformer>> * string -> Microsoft.ML.Transforms.Onnx.DnnImageFeaturizerEstimator
<Extension()>
Public Function DnnFeaturizeImage (catalog As TransformsCatalog, outputColumnName As String, modelFactory As Func(Of DnnImageFeaturizerInput, EstimatorChain(Of ColumnCopyingTransformer)), Optional inputColumnName As String = Nothing) As DnnImageFeaturizerEstimator

Parameters

catalog
TransformsCatalog

The transform's catalog.

outputColumnName
String

The name of the column resulting from the transformation of inputColumnName.

modelFactory
Func<DnnImageFeaturizerInput,EstimatorChain<ColumnCopyingTransformer>>

An extension method on the DnnImageModelSelector that creates a chain of two OnnxScoringEstimator (one for preprocessing and one with a pretrained image DNN) with specific models included in a package together with that extension method.

inputColumnName
String

Name of column to transform. If set to null, the value of the outputColumnName will be used as source.

Returns

Examples

using System.IO;
using System.Linq;
using Microsoft.ML;
using Microsoft.ML.Data;

namespace Samples.Dynamic
{
    public static class DnnFeaturizeImage
    {
        public static void Example()
        {
            // Create a new ML context, for ML.NET operations. It can be used for
            // exception tracking and logging, as well as the source of randomness.
            var mlContext = new MLContext();

            // Downloading a few images, and an images.tsv file, which contains a
            // list of the files from the dotnet/machinelearning/test/data/images/.
            // If you inspect the fileSystem, after running this line, an "images"
            // folder will be created, containing 4 images, and a .tsv file
            // enumerating the images. 
            var imagesDataFile = Microsoft.ML.SamplesUtils.DatasetUtils
                .GetSampleImages();

            // Preview of the content of the images.tsv file, which lists the images
            // to operate on
            //
            // imagePath    imageType
            // tomato.bmp   tomato
            // banana.jpg   banana
            // hotdog.jpg   hotdog
            // tomato.jpg   tomato

            var data = mlContext.Data.CreateTextLoader(new TextLoader.Options()
            {
                Columns = new[]
                {
                        new TextLoader.Column("ImagePath", DataKind.String, 0),
                        new TextLoader.Column("Name", DataKind.String, 1),
                }
            }).Load(imagesDataFile);

            var imagesFolder = Path.GetDirectoryName(imagesDataFile);

            // Installing the Microsoft.ML.DNNImageFeaturizer packages copies the models in the
            // `DnnImageModels` folder. 
            // Image loading pipeline. 
            var pipeline = mlContext.Transforms.LoadImages("ImageObject",
                imagesFolder, "ImagePath")
                .Append(mlContext.Transforms.ResizeImages("ImageObject", imageWidth:
                    224, imageHeight: 224))
                .Append(mlContext.Transforms.ExtractPixels("Pixels", "ImageObject"))
                .Append(mlContext.Transforms.DnnFeaturizeImage("FeaturizedImage",
                    m => m.ModelSelector.ResNet18(mlContext, m.OutputColumn, m
                    .InputColumn), "Pixels"));

            var transformedData = pipeline.Fit(data).Transform(data);

            var FeaturizedImageColumnsPerRow = transformedData.GetColumn<float[]>(
                "FeaturizedImage").ToArray();

            // Preview of FeaturizedImageColumnsPerRow for the first row,
            // FeaturizedImageColumnsPerRow[0]
            //
            // 0.696136236
            // 0.2661711
            // 0.440882325
            // 0.157903448
            // 0.0339231342
            // 0
            // 0.0936501548
            // 0.159010679
            // 0.394427955

        }
    }
}

Applies to