Develop an object detection client application

After you’ve trained an object detection model, you can use the Azure AI Custom Vision SDK to develop a client application that submits new images to be analyzed.

C#Copy

using System;
using System.IO;
using Microsoft.Azure.CognitiveServices.Vision.CustomVision.Prediction;

// Authenticate a client for the prediction API
CustomVisionPredictionClient prediction_client = new CustomVisionPredictionClient(new ApiKeyServiceClientCredentials("<YOUR_PREDICTION_RESOURCE_KEY>"))
{
    Endpoint = "<YOUR_PREDICTION_RESOURCE_ENDPOINT>"
};

// Get classification predictions for an image
MemoryStream image_data = new MemoryStream(File.ReadAllBytes("<PATH_TO_IMAGE_FILE>"));
var result = prediction_client.DetectImage("<YOUR_PROJECT_ID>",
                                             "<YOUR_PUBLISHED_MODEL_NAME>",
                                             image_data);

// Process predictions
foreach (var prediction in result.Predictions)
{
    if (prediction.Probability > 0.5)
    {
        var left = prediction.BoundingBox.Left;
        var top = prediction.BoundingBox.Top;
        var height = prediction.BoundingBox.Height;
        var width =  prediction.BoundingBox.Width;
        Console.WriteLine($"{prediction.TagName} ({prediction.Probability})");
        Console.WriteLine($"  Left:{left}, Top:{top}, Height:{height}, Width:{width}");
    }
}

contact