Method: Aws::Rekognition::Client#compare_faces

Defined in:: lib/aws-sdk-rekognition/client.rb

#compare_faces(params = {}) ⇒ `Types::CompareFacesResponse`

Compares a face in the source input image with each of the 100 largest faces detected in the target input image.

If the source image contains multiple faces, the service detects the largest face and compares it with each face detected in the target image.

<note markdown=“1”> CompareFaces uses machine learning algorithms, which are probabilistic. A false negative is an incorrect prediction that a face in the target image has a low similarity confidence score when compared to the face in the source image. To reduce the probability of false negatives, we recommend that you compare the target image against multiple source images. If you plan to use ‘CompareFaces` to make a decision that impacts an individual’s rights, privacy, or access to services, we recommend that you pass the result to a human for review and further validation before taking action.

</note>

You pass the input and target images either as base64-encoded image bytes or as references to images in an Amazon S3 bucket. If you use the AWS CLI to call Amazon Rekognition operations, passing image bytes isn’t supported. The image must be formatted as a PNG or JPEG file.

In response, the operation returns an array of face matches ordered by similarity score in descending order. For each face match, the response provides a bounding box of the face, facial landmarks, pose details (pitch, roll, and yaw), quality (brightness and sharpness), and confidence value (indicating the level of confidence that the bounding box contains a face). The response also provides a similarity score, which indicates how closely the faces match.

<note markdown=“1”> By default, only faces with a similarity score of greater than or equal to 80% are returned in the response. You can change this value by specifying the ‘SimilarityThreshold` parameter.

</note>

‘CompareFaces` also returns an array of faces that don’t match the source image. For each face, it returns a bounding box, confidence value, landmarks, pose details, and quality. The response also returns information about the face in the source image, including the bounding box of the face and confidence value.

The ‘QualityFilter` input parameter allows you to filter out detected faces that don’t meet a required quality bar. The quality bar is based on a variety of common use cases. Use `QualityFilter` to set the quality bar by specifying `LOW`, `MEDIUM`, or `HIGH`. If you do not want to filter detected faces, specify `NONE`. The default value is `NONE`.

If the image doesn’t contain Exif metadata, ‘CompareFaces` returns orientation information for the source and target images. Use these values to display the images with the correct image orientation.

If no faces are detected in the source or target images, ‘CompareFaces` returns an `InvalidParameterException` error.

<note markdown=“1”> This is a stateless API operation. That is, data returned by this operation doesn’t persist.

</note>

For an example, see Comparing Faces in Images in the Amazon Rekognition Developer Guide.

This operation requires permissions to perform the ‘rekognition:CompareFaces` action.

Examples:

Example: To compare two images


# This operation compares the largest face detected in the source image with each face detected in the target image.

resp = client.compare_faces({
  similarity_threshold: 90, 
  source_image: {
    s3_object: {
      bucket: "mybucket", 
      name: "mysourceimage", 
    }, 
  }, 
  target_image: {
    s3_object: {
      bucket: "mybucket", 
      name: "mytargetimage", 
    }, 
  }, 
})

resp.to_h outputs the following:
{
  face_matches: [
    {
      face: {
        bounding_box: {
          height: 0.33481481671333313, 
          left: 0.31888890266418457, 
          top: 0.4933333396911621, 
          width: 0.25, 
        }, 
        confidence: 99.9991226196289, 
      }, 
      similarity: 100, 
    }, 
  ], 
  source_image_face: {
    bounding_box: {
      height: 0.33481481671333313, 
      left: 0.31888890266418457, 
      top: 0.4933333396911621, 
      width: 0.25, 
    }, 
    confidence: 99.9991226196289, 
  }, 
}

Request syntax with placeholder values


resp = client.compare_faces({
  source_image: { # required
    bytes: "data",
    s3_object: {
      bucket: "S3Bucket",
      name: "S3ObjectName",
      version: "S3ObjectVersion",
    },
  },
  target_image: { # required
    bytes: "data",
    s3_object: {
      bucket: "S3Bucket",
      name: "S3ObjectName",
      version: "S3ObjectVersion",
    },
  },
  similarity_threshold: 1.0,
  quality_filter: "NONE", # accepts NONE, AUTO, LOW, MEDIUM, HIGH
})

Response structure


resp.source_image_face.bounding_box.width #=> Float
resp.source_image_face.bounding_box.height #=> Float
resp.source_image_face.bounding_box.left #=> Float
resp.source_image_face.bounding_box.top #=> Float
resp.source_image_face.confidence #=> Float
resp.face_matches #=> Array
resp.face_matches[0].similarity #=> Float
resp.face_matches[0].face.bounding_box.width #=> Float
resp.face_matches[0].face.bounding_box.height #=> Float
resp.face_matches[0].face.bounding_box.left #=> Float
resp.face_matches[0].face.bounding_box.top #=> Float
resp.face_matches[0].face.confidence #=> Float
resp.face_matches[0].face.landmarks #=> Array
resp.face_matches[0].face.landmarks[0].type #=> String, one of "eyeLeft", "eyeRight", "nose", "mouthLeft", "mouthRight", "leftEyeBrowLeft", "leftEyeBrowRight", "leftEyeBrowUp", "rightEyeBrowLeft", "rightEyeBrowRight", "rightEyeBrowUp", "leftEyeLeft", "leftEyeRight", "leftEyeUp", "leftEyeDown", "rightEyeLeft", "rightEyeRight", "rightEyeUp", "rightEyeDown", "noseLeft", "noseRight", "mouthUp", "mouthDown", "leftPupil", "rightPupil", "upperJawlineLeft", "midJawlineLeft", "chinBottom", "midJawlineRight", "upperJawlineRight"
resp.face_matches[0].face.landmarks[0].x #=> Float
resp.face_matches[0].face.landmarks[0].y #=> Float
resp.face_matches[0].face.pose.roll #=> Float
resp.face_matches[0].face.pose.yaw #=> Float
resp.face_matches[0].face.pose.pitch #=> Float
resp.face_matches[0].face.quality.brightness #=> Float
resp.face_matches[0].face.quality.sharpness #=> Float
resp.face_matches[0].face.emotions #=> Array
resp.face_matches[0].face.emotions[0].type #=> String, one of "HAPPY", "SAD", "ANGRY", "CONFUSED", "DISGUSTED", "SURPRISED", "CALM", "UNKNOWN", "FEAR"
resp.face_matches[0].face.emotions[0].confidence #=> Float
resp.face_matches[0].face.smile.value #=> Boolean
resp.face_matches[0].face.smile.confidence #=> Float
resp.unmatched_faces #=> Array
resp.unmatched_faces[0].bounding_box.width #=> Float
resp.unmatched_faces[0].bounding_box.height #=> Float
resp.unmatched_faces[0].bounding_box.left #=> Float
resp.unmatched_faces[0].bounding_box.top #=> Float
resp.unmatched_faces[0].confidence #=> Float
resp.unmatched_faces[0].landmarks #=> Array
resp.unmatched_faces[0].landmarks[0].type #=> String, one of "eyeLeft", "eyeRight", "nose", "mouthLeft", "mouthRight", "leftEyeBrowLeft", "leftEyeBrowRight", "leftEyeBrowUp", "rightEyeBrowLeft", "rightEyeBrowRight", "rightEyeBrowUp", "leftEyeLeft", "leftEyeRight", "leftEyeUp", "leftEyeDown", "rightEyeLeft", "rightEyeRight", "rightEyeUp", "rightEyeDown", "noseLeft", "noseRight", "mouthUp", "mouthDown", "leftPupil", "rightPupil", "upperJawlineLeft", "midJawlineLeft", "chinBottom", "midJawlineRight", "upperJawlineRight"
resp.unmatched_faces[0].landmarks[0].x #=> Float
resp.unmatched_faces[0].landmarks[0].y #=> Float
resp.unmatched_faces[0].pose.roll #=> Float
resp.unmatched_faces[0].pose.yaw #=> Float
resp.unmatched_faces[0].pose.pitch #=> Float
resp.unmatched_faces[0].quality.brightness #=> Float
resp.unmatched_faces[0].quality.sharpness #=> Float
resp.unmatched_faces[0].emotions #=> Array
resp.unmatched_faces[0].emotions[0].type #=> String, one of "HAPPY", "SAD", "ANGRY", "CONFUSED", "DISGUSTED", "SURPRISED", "CALM", "UNKNOWN", "FEAR"
resp.unmatched_faces[0].emotions[0].confidence #=> Float
resp.unmatched_faces[0].smile.value #=> Boolean
resp.unmatched_faces[0].smile.confidence #=> Float
resp.source_image_orientation_correction #=> String, one of "ROTATE_0", "ROTATE_90", "ROTATE_180", "ROTATE_270"
resp.target_image_orientation_correction #=> String, one of "ROTATE_0", "ROTATE_90", "ROTATE_180", "ROTATE_270"

Parameters:

params (Hash) (defaults to: {}) —

({})

Options Hash (params):

:source_image (required, Types::Image) —

The input image as base64-encoded bytes or an S3 object. If you use the AWS CLI to call Amazon Rekognition operations, passing base64-encoded image bytes is not supported.

If you are using an AWS SDK to call Amazon Rekognition, you might not need to base64-encode image bytes passed using the ‘Bytes` field. For more information, see Images in the Amazon Rekognition developer guide.
:target_image (required, Types::Image) —

The target image as base64-encoded bytes or an S3 object. If you use the AWS CLI to call Amazon Rekognition operations, passing base64-encoded image bytes is not supported.

If you are using an AWS SDK to call Amazon Rekognition, you might not need to base64-encode image bytes passed using the ‘Bytes` field. For more information, see Images in the Amazon Rekognition developer guide.
:similarity_threshold (Float) —

The minimum level of confidence in the face matches that a match must meet to be included in the ‘FaceMatches` array.
:quality_filter (String) —

A filter that specifies a quality bar for how much filtering is done to identify faces. Filtered faces aren’t compared. If you specify ‘AUTO`, Amazon Rekognition chooses the quality bar. If you specify `LOW`, `MEDIUM`, or `HIGH`, filtering removes all faces that don’t meet the chosen quality bar. The quality bar is based on a variety of common use cases. Low-quality detections can occur for a number of reasons. Some examples are an object that’s misidentified as a face, a face that’s too blurry, or a face with a pose that’s too extreme to use. If you specify ‘NONE`, no filtering is performed. The default value is `NONE`.

To use quality filtering, the collection you are using must be associated with version 3 of the face model or higher.

Returns:

(Types::CompareFacesResponse) —
Returns a response object which responds to the following methods:
- #source_image_face => Types::ComparedSourceImageFace
- #face_matches => Array<Types::CompareFacesMatch>
- #unmatched_faces => Array<Types::ComparedFace>
- #source_image_orientation_correction => String
- #target_image_orientation_correction => String

# File 'lib/aws-sdk-rekognition/client.rb', line 865

def compare_faces(params = {}, options = {})
  req = build_request(:compare_faces, params)
  req.send_request(options)
end