Google.Cloud.Vision.V1

Google.Cloud.Vision.V1 is a.NET client library for the Google Cloud Vision API.

Note: This documentation is for version 2.3.0 of the library. Some samples may not work with other versions.

Installation

Install the Google.Cloud.Vision.V1 package from NuGet. Add it to your project in the normal way (for example by right-clicking on the project in Visual Studio and choosing "Manage NuGet Packages...").

Authentication

When running on Google Cloud Platform, no action needs to be taken to authenticate.

Otherwise, the simplest way of authenticating your API calls is to download a service account JSON file then set the GOOGLE_APPLICATION_CREDENTIALS environment variable to refer to it. The credentials will automatically be used to authenticate. See the Getting Started With Authentication guide for more details.

Getting started

All operations are performed through the following client classes:

Create a client instance by calling the static Create or CreateAsync methods. Alternatively, use the builder class associated with each client class (e.g. ImageAnnotatorClientBuilder for ImageAnnotatorClient) as an easy way of specifying custom credentials, settings, or a custom endpoint. Clients are thread-safe, and we recommend using a single instance across your entire application unless you have a particular need to configure multiple client objects separately.

The "core" method BatchAnnotateImages can perform multiple (potentially different) annotations on multiple images, but convenience methods are provided for common cases of working with a single image, and for performing a single annotation operation on a single image.

Sample code

Constructing an Image object

There are various factory methods on the Image class to allow instances to be constructed from files, streams, byte arrays and URIs.

Image image1 = Image.FromFile("Pictures/LocalImage.jpg");
// Fetched locally by the client, then uploaded to the server
Image image2 = Image.FetchFromUri("https://cloud.google.com/images/devtools-icon-64x64.png");
// Fetched by the Google Cloud Vision server
Image image3 = Image.FromUri("https://cloud.google.com/images/devtools-icon-64x64.png");
// Google Cloud Storage URI
Image image4 = Image.FromUri("gs://my-bucket/my-file");

byte[] bytes = ReadImageData(); // For example, from a database
Image image5 = Image.FromBytes(bytes);

using (Stream stream = OpenImageStream()) // Any regular .NET stream
{
    Image image6 = Image.FromStream(stream);
}

All IO-related methods have async equivalents.

Detect faces in a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
IReadOnlyList<FaceAnnotation> result = client.DetectFaces(image);
foreach (FaceAnnotation face in result)
{
    string poly = string.Join(" - ", face.BoundingPoly.Vertices.Select(v => $"({v.X}, {v.Y})"));
    Console.WriteLine($"Confidence: {(int)(face.DetectionConfidence * 100)}%; BoundingPoly: {poly}");
}

Detect text in a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
IReadOnlyList<EntityAnnotation> textAnnotations = client.DetectText(image);
foreach (EntityAnnotation text in textAnnotations)
{
    Console.WriteLine($"Description: {text.Description}");
}

Detect document text in a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
TextAnnotation text = client.DetectDocumentText(image);
Console.WriteLine($"Text: {text.Text}");
foreach (var page in text.Pages)
{
    foreach (var block in page.Blocks)
    {
        string box = string.Join(" - ", block.BoundingBox.Vertices.Select(v => $"({v.X}, {v.Y})"));
        Console.WriteLine($"Block {block.BlockType} at {box}");
        foreach (var paragraph in block.Paragraphs)
        {
            box = string.Join(" - ", paragraph.BoundingBox.Vertices.Select(v => $"({v.X}, {v.Y})"));
            Console.WriteLine($"  Paragraph at {box}");
            foreach (var word in paragraph.Words)
            {
                Console.WriteLine($"    Word: {string.Join("", word.Symbols.Select(s => s.Text))}");
            }
        }
    }
}

Detect labels in a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
IReadOnlyList<EntityAnnotation> labels = client.DetectLabels(image);
foreach (EntityAnnotation label in labels)
{
    Console.WriteLine($"Score: {(int)(label.Score * 100)}%; Description: {label.Description}");
}

Detect landmarks in a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
IReadOnlyList<EntityAnnotation> result = client.DetectLandmarks(image);
foreach (EntityAnnotation landmark in result)
{
    Console.WriteLine($"Score: {(int)(landmark.Score * 100)}%; Description: {landmark.Description}");
}

Detect logos in a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
IReadOnlyList<EntityAnnotation> logos = client.DetectLogos(image);
foreach (EntityAnnotation logo in logos)
{
    Console.WriteLine($"Description: {logo.Description}");
}

Perform "safe search" processing on a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
SafeSearchAnnotation annotation = client.DetectSafeSearch(image);
// Each category is classified as Very Unlikely, Unlikely, Possible, Likely or Very Likely.
Console.WriteLine($"Adult? {annotation.Adult}");
Console.WriteLine($"Spoof? {annotation.Spoof}");
Console.WriteLine($"Violence? {annotation.Violence}");
Console.WriteLine($"Medical? {annotation.Medical}");

Perform image property processing on a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
ImageProperties properties = client.DetectImageProperties(image);
ColorInfo dominantColor = properties.DominantColors.Colors.OrderByDescending(c => c.PixelFraction).First();
Console.WriteLine($"Dominant color in image: {dominantColor}");

Suggest crop hints for a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
CropHintsAnnotation cropHints = client.DetectCropHints(image);
foreach (CropHint hint in cropHints.CropHints)
{
    Console.WriteLine("Crop hint:");
    string poly = string.Join(" - ", hint.BoundingPoly.Vertices.Select(v => $"({v.X}, {v.Y})"));
    Console.WriteLine($"  Poly: {poly}");
    Console.WriteLine($"  Confidence: {hint.Confidence}");
    Console.WriteLine($"  Importance fraction: {hint.ImportanceFraction}");
}

Perform analysis for other web references on a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
WebDetection webDetection = client.DetectWebInformation(image);
foreach (WebDetection.Types.WebImage webImage in webDetection.FullMatchingImages)
{
    Console.WriteLine($"Full image: {webImage.Url} ({webImage.Score})");
}
foreach (WebDetection.Types.WebImage webImage in webDetection.PartialMatchingImages)
{
    Console.WriteLine($"Partial image: {webImage.Url} ({webImage.Score})");
}
foreach (WebDetection.Types.WebPage webPage in webDetection.PagesWithMatchingImages)
{
    Console.WriteLine($"Page with matching image: {webPage.Url} ({webPage.Score})");
}
foreach (WebDetection.Types.WebEntity entity in webDetection.WebEntities)
{
    Console.WriteLine($"Web entity: {entity.EntityId} / {entity.Description} ({entity.Score})");
}

Detect localized objects in a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
IReadOnlyList<LocalizedObjectAnnotation> annotations = client.DetectLocalizedObjects(image);
foreach (LocalizedObjectAnnotation annotation in annotations)
{
    string poly = string.Join(" - ", annotation.BoundingPoly.NormalizedVertices.Select(v => $"({v.X}, {v.Y})"));
    Console.WriteLine(
        $"Name: {annotation.Name}; ID: {annotation.Mid}; Score: {annotation.Score}; Bounding poly: {poly}");
}

Detect faces and landmarks in a single image

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
AnnotateImageRequest request = new AnnotateImageRequest
{
    Image = image,
    Features =
    {
        new Feature { Type = Feature.Types.Type.FaceDetection },
        // By default, no limits are put on the number of results per annotation.
        // Use the MaxResults property to specify a limit.
        new Feature { Type = Feature.Types.Type.LandmarkDetection, MaxResults = 5 },
    }
};
AnnotateImageResponse response = client.Annotate(request);
Console.WriteLine("Faces:");
foreach (FaceAnnotation face in response.FaceAnnotations)
{
    string poly = string.Join(" - ", face.BoundingPoly.Vertices.Select(v => $"({v.X}, {v.Y})"));
    Console.WriteLine($"  Confidence: {(int)(face.DetectionConfidence * 100)}%; BoundingPoly: {poly}");
}
Console.WriteLine("Landmarks:");
foreach (EntityAnnotation landmark in response.LandmarkAnnotations)
{
    Console.WriteLine($"Score: {(int)(landmark.Score * 100)}%; Description: {landmark.Description}");
}
if (response.Error != null)
{
    Console.WriteLine($"Error detected: {response.Error}");
}

Detect faces in one image and logos in another

ImageAnnotatorClient client = ImageAnnotatorClient.Create();
// Perform face recognition on one image, and logo recognition on another.
AnnotateImageRequest request1 = new AnnotateImageRequest
{
    Image = image1,
    Features = { new Feature { Type = Feature.Types.Type.FaceDetection } }
};
AnnotateImageRequest request2 = new AnnotateImageRequest
{
    Image = image2,
    Features = { new Feature { Type = Feature.Types.Type.LogoDetection } }
};

BatchAnnotateImagesResponse response = client.BatchAnnotateImages(new[] { request1, request2 });
Console.WriteLine("Faces in image 1:");
foreach (FaceAnnotation face in response.Responses[0].FaceAnnotations)
{
    string poly = string.Join(" - ", face.BoundingPoly.Vertices.Select(v => $"({v.X}, {v.Y})"));
    Console.WriteLine($"  Confidence: {(int)(face.DetectionConfidence * 100)}%; BoundingPoly: {poly}");
}
Console.WriteLine("Logos in image 2:");
foreach (EntityAnnotation logo in response.Responses[1].LogoAnnotations)
{
    Console.WriteLine($"Description: {logo.Description}");
}
foreach (Status error in response.Responses.Select(r => r.Error))
{
    Console.WriteLine($"Error detected: error");
}

After creating and populating a product set, the products can be detected within images.

ProductSetName productSetName = new ProductSetName(projectId, locationId, productSetId);
ImageAnnotatorClient client = ImageAnnotatorClient.Create();
ProductSearchParams searchParams = new ProductSearchParams
{
    ProductCategories = { "apparel" },
    ProductSetAsProductSetName = productSetName,
};
ProductSearchResults results = client.DetectSimilarProducts(image, searchParams);
foreach (var result in results.Results)
{
    Console.WriteLine($"{result.Product.DisplayName}: {result.Score}");
}

A filter can be applied to the search, to match only products with specific labels.

ProductSetName productSetName = new ProductSetName(projectId, locationId, productSetId);
ImageAnnotatorClient client = ImageAnnotatorClient.Create();
ProductSearchParams searchParams = new ProductSearchParams
{
    ProductCategories = { "apparel" },
    ProductSetAsProductSetName = productSetName,
    Filter = "style=womens"
};
ProductSearchResults results = client.DetectSimilarProducts(image, searchParams);
foreach (var result in results.Results)
{
    Console.WriteLine($"{result.Product.DisplayName}: {result.Score}");
}

Error handling

All the methods which annotate a single image (and therefore have a single response) throw AnnotateImageException if the response contains an error.

// We create a request which passes simple validation, but isn't a valid image.
Image image = Image.FromBytes(new byte[10]);
ImageAnnotatorClient client = ImageAnnotatorClient.Create();
try
{
    IReadOnlyList<EntityAnnotation> logos = client.DetectLogos(image);
    // Normally use logos here...
}
catch (AnnotateImageException e)
{
    AnnotateImageResponse response = e.Response;
    Console.WriteLine(response.Error);
}

The BatchAnnotateImages method does not throw this exception, but BatchAnnotateImagesResponse.ThrowOnAnyError() checks all responses are successful, throwing an AggregateException if there are any errors. The AggregateException contains one AnnotateImageException for each response that contains an error.

// We create a request which passes simple validation, but isn't a valid image.
Image image = Image.FromBytes(new byte[10]);
// Just a single request in this example, but usually BatchAnnotateImages would be
// used with multiple requests.
var request = new AnnotateImageRequest
{
    Image = image,
    Features = { new Feature { Type = Feature.Types.Type.SafeSearchDetection } }
};
ImageAnnotatorClient client = ImageAnnotatorClient.Create();
try
{
    BatchAnnotateImagesResponse response = client.BatchAnnotateImages(new[] { request });
    // ThrowOnAnyError will throw if any individual response in response.Responses
    // contains an error. Other responses may still have useful results.
    // Errors can be detected manually by checking the Error property in each
    // individual response.
    response.ThrowOnAnyError();
}
catch (AggregateException e)
{
    // Because a batch can have multiple errors, the exception thrown is AggregateException.
    // Each inner exception is an AnnotateImageException
    foreach (AnnotateImageException innerException in e.InnerExceptions)
    {
        Console.WriteLine(innerException.Response.Error);
    }
}