Vision API

vision/face-detection/src/main/java/com/google/cloud/vision/samples/facedetect/FaceDetectApp.java

View on GitHub


using Google.Cloud.Vision.V1;
using System;
using System.Linq;

var client = ImageAnnotatorClient.Create();

Java

View on GitHub

import com.google.api.client.googleapis.auth.oauth2.GoogleCredential;
import com.google.api.client.googleapis.javanet.GoogleNetHttpTransport;
import com.google.api.client.json.JsonFactory;
import com.google.api.client.json.jackson2.JacksonFactory;
import com.google.api.services.vision.v1.Vision;
import com.google.api.services.vision.v1.VisionScopes;
import com.google.api.services.vision.v1.model.AnnotateImageRequest;
import com.google.api.services.vision.v1.model.AnnotateImageResponse;
import com.google.api.services.vision.v1.model.BatchAnnotateImagesRequest;
import com.google.api.services.vision.v1.model.BatchAnnotateImagesResponse;
import com.google.api.services.vision.v1.model.FaceAnnotation;
import com.google.api.services.vision.v1.model.Feature;
import com.google.api.services.vision.v1.model.Image;
import com.google.api.services.vision.v1.model.Vertex;
import com.google.common.collect.ImmutableList;
import java.awt.BasicStroke;
import java.awt.Color;
import java.awt.Graphics2D;
import java.awt.Polygon;
import java.awt.image.BufferedImage;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.security.GeneralSecurityException;
import java.util.List;
import javax.imageio.ImageIO;

/** Connects to the Vision API using Application Default Credentials. */
public static Vision getVisionService() throws IOException, GeneralSecurityException {
  GoogleCredential credential =
      GoogleCredential.getApplicationDefault().createScoped(VisionScopes.all());
  JsonFactory jsonFactory = JacksonFactory.getDefaultInstance();
  return new Vision.Builder(GoogleNetHttpTransport.newTrustedTransport(), jsonFactory, credential)
      .setApplicationName(APPLICATION_NAME)
      .build();
}

Node.js

View on GitHub

// By default, the client will authenticate using the service account file
// specified by the GOOGLE_APPLICATION_CREDENTIALS environment variable and use
// the project specified by the GCLOUD_PROJECT environment variable. See
// https://googlecloudplatform.github.io/gcloud-node/#/docs/google-cloud/latest/guides/authentication
const vision = require('@google-cloud/vision');

// Creates a client
const client = new vision.ImageAnnotatorClient();

const fs = require('fs');

PHP

View on GitHub

use Google\Cloud\Vision\V1\ImageAnnotatorClient;

$imageAnnotator = new ImageAnnotatorClient();

Python

View on GitHub

from google.cloud import vision
from google.cloud.vision import types
from PIL import Image, ImageDraw

client = vision.ImageAnnotatorClient()

Ruby

View on GitHub

require "google/cloud/vision"

image_annotator = Google::Cloud::Vision::ImageAnnotator.new

Send a face detection request

To construct a request to the Vision API, first consult the API documentation. In this case, you'll be asking the images resource to annotate your image. A request to this API takes the form of an object with a requests list. Each item in this list contains two bits of information:

The base64-encoded image data
A list of features you'd like annotated about that image.

For this example, you'll simply request FACE_DETECTION annotation on one image, and return the relevant portion of the response:

C#

vision/face-detection/src/main/java/com/google/cloud/vision/samples/facedetect/FaceDetectApp.java

View on GitHub

var response = client.DetectFaces(Image.FromFile(args[0]));

Java

View on GitHub

/** Gets up to {@code maxResults} faces for an image stored at {@code path}. */
public List<FaceAnnotation> detectFaces(Path path, int maxResults) throws IOException {
  byte[] data = Files.readAllBytes(path);

  AnnotateImageRequest request =
      new AnnotateImageRequest()
          .setImage(new Image().encodeContent(data))
          .setFeatures(
              ImmutableList.of(
                  new Feature().setType("FACE_DETECTION").setMaxResults(maxResults)));
  Vision.Images.Annotate annotate =
      vision
          .images()
          .annotate(new BatchAnnotateImagesRequest().setRequests(ImmutableList.of(request)));
  // Due to a bug: requests to Vision API containing large images fail when GZipped.
  annotate.setDisableGZipContent(true);

  BatchAnnotateImagesResponse batchResponse = annotate.execute();
  assert batchResponse.getResponses().size() == 1;
  AnnotateImageResponse response = batchResponse.getResponses().get(0);
  if (response.getFaceAnnotations() == null) {
    throw new IOException(
        response.getError() != null
            ? response.getError().getMessage()
            : "Unknown error getting image annotations");
  }
  return response.getFaceAnnotations();
}

Node.js

View on GitHub

async function detectFaces(inputFile) {
  // Make a call to the Vision API to detect the faces
  const request = {image: {source: {filename: inputFile}}};
  const results = await client.faceDetection(request);
  const faces = results[0].faceAnnotations;
  const numFaces = faces.length;
  console.log(`Found ${numFaces} face${numFaces === 1 ? '' : 's'}.`);
  return faces;
}

PHP

View on GitHub

# annotate the image
// $path = 'path/to/your/image.jpg'
$image = file_get_contents($path);
$response = $imageAnnotator->faceDetection($image);
$faces = $response->getFaceAnnotations();

Python

View on GitHub

def detect_face(face_file, max_results=4):
    """Uses the Vision API to detect faces in the given file.

    Args:
        face_file: A file-like object containing an image with faces.

    Returns:
        An array of Face objects with information about the picture.
    """
    client = vision.ImageAnnotatorClient()

    content = face_file.read()
    image = types.Image(content=content)

    return client.face_detection(
        image=image, max_results=max_results).face_annotations

Ruby

View on GitHub

response = image_annotator.face_detection image: path_to_image_file

Process the response

Congratulations - you've detected the faces in your image! The response to our face annotation request includes a bunch of metadata about the detected faces, which include coordinates of a polygon encompassing the face. At this point, though, this is only a list of numbers. Let's use them to confirm that you have, in fact, found the faces in your image. We'll draw polygons onto a copy of the image, using the coordinates returned by the Vision API:

C#

vision/face-detection/src/main/java/com/google/cloud/vision/samples/facedetect/FaceDetectApp.java

View on GitHub

using (var image = System.Drawing.Image.FromFile(args[0]))
using (System.Drawing.Graphics g = System.Drawing.Graphics.FromImage(image))
{
    var cyanPen = new System.Drawing.Pen(System.Drawing.Color.Cyan, 3);
    foreach (var annotation in response)
    {
        g.DrawPolygon(cyanPen, annotation.BoundingPoly.Vertices.Select(
            (vertex) => new System.Drawing.Point(vertex.X, vertex.Y)).ToArray());
        // ...
    }
    // ...
}

Java

View on GitHub

/** Reads image {@code inputPath} and writes {@code outputPath} with {@code faces} outlined. */
private static void writeWithFaces(Path inputPath, Path outputPath, List<FaceAnnotation> faces)
    throws IOException {
  BufferedImage img = ImageIO.read(inputPath.toFile());
  annotateWithFaces(img, faces);
  ImageIO.write(img, "jpg", outputPath.toFile());
}

/** Annotates an image {@code img} with a polygon around each face in {@code faces}. */
public static void annotateWithFaces(BufferedImage img, List<FaceAnnotation> faces) {
  for (FaceAnnotation face : faces) {
    annotateWithFace(img, face);
  }
}

/** Annotates an image {@code img} with a polygon defined by {@code face}. */
private static void annotateWithFace(BufferedImage img, FaceAnnotation face) {
  Graphics2D gfx = img.createGraphics();
  Polygon poly = new Polygon();
  for (Vertex vertex : face.getFdBoundingPoly().getVertices()) {
    poly.addPoint(vertex.getX(), vertex.getY());
  }
  gfx.setStroke(new BasicStroke(5));
  gfx.setColor(new Color(0x00ff00));
  gfx.draw(poly);
}

Node.js

We use the node-canvas library to draw onto images.

View on GitHub

async function highlightFaces(inputFile, faces, outputFile, PImage) {
  // Open the original image
  const stream = fs.createReadStream(inputFile);
  let promise;
  if (inputFile.match(/\.jpg$/)) {
    promise = PImage.decodeJPEGFromStream(stream);
  } else if (inputFile.match(/\.png$/)) {
    promise = PImage.decodePNGFromStream(stream);
  } else {
    throw new Error(`Unknown filename extension ${inputFile}`);
  }
  const img = await promise;
  const context = img.getContext('2d');
  context.drawImage(img, 0, 0, img.width, img.height, 0, 0);

  // Now draw boxes around all the faces
  context.strokeStyle = 'rgba(0,255,0,0.8)';
  context.lineWidth = '5';

  faces.forEach(face => {
    context.beginPath();
    let origX = 0;
    let origY = 0;
    face.boundingPoly.vertices.forEach((bounds, i) => {
      if (i === 0) {
        origX = bounds.x;
        origY = bounds.y;
        context.moveTo(bounds.x, bounds.y);
      } else {
        context.lineTo(bounds.x, bounds.y);
      }
    });
    context.lineTo(origX, origY);
    context.stroke();
  });

  // Write the result to a file
  console.log(`Writing to file ${outputFile}`);
  const writeStream = fs.createWriteStream(outputFile);
  await PImage.encodePNGToStream(img, writeStream);
}

PHP

We use the GD extension to draw onto images.

View on GitHub

# draw box around faces
if ($faces && $outFile) {
    $imageCreateFunc = [
        'png' => 'imagecreatefrompng',
        'gd' => 'imagecreatefromgd',
        'gif' => 'imagecreatefromgif',
        'jpg' => 'imagecreatefromjpeg',
        'jpeg' => 'imagecreatefromjpeg',
    ];
    $imageWriteFunc = [
        'png' => 'imagepng',
        'gd' => 'imagegd',
        'gif' => 'imagegif',
        'jpg' => 'imagejpeg',
        'jpeg' => 'imagejpeg',
    ];

    copy($path, $outFile);
    $ext = strtolower(pathinfo($path, PATHINFO_EXTENSION));
    if (!array_key_exists($ext, $imageCreateFunc)) {
        throw new \Exception('Unsupported image extension');
    }
    $outputImage = call_user_func($imageCreateFunc[$ext], $outFile);

    foreach ($faces as $face) {
        $vertices = $face->getBoundingPoly()->getVertices();
        if ($vertices) {
            $x1 = $vertices[0]->getX();
            $y1 = $vertices[0]->getY();
            $x2 = $vertices[2]->getX();
            $y2 = $vertices[2]->getY();
            imagerectangle($outputImage, $x1, $y1, $x2, $y2, 0x00ff00);
        }
    }

Python

View on GitHub

def highlight_faces(image, faces, output_filename):
    """Draws a polygon around the faces, then saves to output_filename.

    Args:
      image: a file containing the image with the faces.
      faces: a list of faces found in the file. This should be in the format
          returned by the Vision API.
      output_filename: the name of the image file to be created, where the
          faces have polygons drawn around them.
    """
    im = Image.open(image)
    draw = ImageDraw.Draw(im)
    # Sepecify the font-family and the font-size
    for face in faces:
        box = [(vertex.x, vertex.y)
               for vertex in face.bounding_poly.vertices]
        draw.line(box + [box[0]], width=5, fill='#00ff00')
        # Place the confidence value/score of the detected faces above the
        # detection box in the output image
        draw.text(((face.bounding_poly.vertices)[0].x,
                   (face.bounding_poly.vertices)[0].y - 30),
                  str(format(face.detection_confidence, '.3f')) + '%',
                  fill='#FF0000')
    im.save(output_filename)

Ruby

We use the rmagick gem to draw onto images.

View on GitHub

require "rmagick"
  response.responses.each do |res|
    res.face_annotations.each do |annotation|
      puts "Face bounds:"
      annotation.bounding_poly.vertices.each do |vertex|
        puts "(#{vertex.x}, #{vertex.y})"
      end

      x1 = annotation.bounding_poly.vertices[0].x.to_i
      y1 = annotation.bounding_poly.vertices[0].y.to_i
      x2 = annotation.bounding_poly.vertices[2].x.to_i
      y2 = annotation.bounding_poly.vertices[2].y.to_i

      photo = Magick::Image.read(path_to_image_file).first
      draw = Magick::Draw.new
      draw.stroke = "green"
      draw.stroke_width 5
      draw.fill_opacity 0
      draw.rectangle x1, y1, x2, y2
      draw.draw photo

      photo.write path_to_output_file
    end
  end

  puts "Output file: #{path_to_output_file}"

Put it all together

C#

vision/face-detection/src/main/java/com/google/cloud/vision/samples/facedetect/FaceDetectApp.java

View on GitHub

static readonly string s_usage = @"dotnet run image-file

Use the Google Cloud Vision API to detect faces in the image.
Writes an output file called image-file.faces.
";
public static void Main(string[] args)
{
    if (args.Length < 1)
    {
        Console.WriteLine(s_usage);
        return;
    }

    var client = ImageAnnotatorClient.Create();
    var response = client.DetectFaces(Image.FromFile(args[0]));

    int numberOfFacesFound = 0;
    using (var image = System.Drawing.Image.FromFile(args[0]))
    using (System.Drawing.Graphics g = System.Drawing.Graphics.FromImage(image))
    {
        var cyanPen = new System.Drawing.Pen(System.Drawing.Color.Cyan, 3);
        foreach (var annotation in response)
        {
            g.DrawPolygon(cyanPen, annotation.BoundingPoly.Vertices.Select(
                (vertex) => new System.Drawing.Point(vertex.X, vertex.Y)).ToArray());
            // ...
        }
        // ...
    }
    // ...
}

To build the sample, open the Solution file dotnet-doc-samples/vision/api/Vision.sln in Visual Studio and build the solution.

To run the sample:

C:\...\> cd dotnet-docs-samples\vision\api\DetectFaces\bin\Debug
C:\...\bin\Debug> DetectFaces ..\..\..\VisionTest\data\face.png

Java

View on GitHub

/** Annotates an image using the Vision API. */
public static void main(String[] args) throws IOException, GeneralSecurityException {
  if (args.length != 2) {
    System.err.println("Usage:");
    System.err.printf(
        "\tjava %s inputImagePath outputImagePath\n", FaceDetectApp.class.getCanonicalName());
    System.exit(1);
  }
  Path inputPath = Paths.get(args[0]);
  Path outputPath = Paths.get(args[1]);
  if (!outputPath.toString().toLowerCase().endsWith(".jpg")) {
    System.err.println("outputImagePath must have the file extension 'jpg'.");
    System.exit(1);
  }

  FaceDetectApp app = new FaceDetectApp(getVisionService());
  List<FaceAnnotation> faces = app.detectFaces(inputPath, MAX_RESULTS);
  System.out.printf("Found %d face%s\n", faces.size(), faces.size() == 1 ? "" : "s");
  System.out.printf("Writing to file %s\n", outputPath);
  app.writeWithFaces(inputPath, outputPath, faces);
}

...

To build and run the sample, run the following commands from the sample code directory:

mvn clean compile assembly:single
java -cp target/vision-face-detection-1.0-SNAPSHOT-jar-with-dependencies.jar \
    com.google.cloud.vision.samples.facedetect.FaceDetectApp \
    data/face.jpg \
    output.jpg

Node.js

View on GitHub

async function main(inputFile, outputFile) {
  const PImage = require('pureimage');
  outputFile = outputFile || 'out.png';
  const faces = await detectFaces(inputFile);
  console.log('Highlighting...');
  await highlightFaces(inputFile, faces, outputFile, PImage);
  console.log('Finished!');
}

To run the sample, run the following command from the sample code directory:

node faceDetection resources/face.png

PHP

View on GitHub

call_user_func($imageWriteFunc[$ext], $outputImage, $outFile);
printf('Output image written to %s' . PHP_EOL, $outFile);

To run the sample, run the following command from the sample code directory:

composer install
php vision.php face test/data/face.png output-image.png

Python

View on GitHub

def main(input_filename, output_filename, max_results):
    with open(input_filename, 'rb') as image:
        faces = detect_face(image, max_results)
        print('Found {} face{}'.format(
            len(faces), '' if len(faces) == 1 else 's'))

        print('Writing to file {}'.format(output_filename))
        # Reset the file pointer, so we can read the file again
        image.seek(0)
        highlight_faces(image, faces, output_filename)

Ruby