AWS Rekognition

Amazon Rekognition makes it easy to add image and video analysis to applications with highly scalable, deep learning technology that requires no machine learning expertise to use. With Amazon Rekognition, you can identify objects, people, text, scenes, and activities in images and videos, as well as detect any inappropriate content. Amazon Rekognition also provides highly accurate facial analysis and facial search capabilities that you can use to detect, analyze, and compare faces for a wide variety of user verification, people counting, and public safety use cases.

Rekognition resource

Sigma IDE currently supports the following image based analytic operations via AWS Rekognition resource component.

  • Compare Faces
  • Detect Faces
  • Detect Labels
  • Detect Text
  • Recognize Celebrities

Rekognition resource operations

To use Rekognition operations from a Lambda function, first drag a Rekognition resource from the resources panel and drop on the required line of the lambda code editor. Then on the configuration panel, select the desired Operation.

Compare Faces

Compare Faces operation can be used to compare a face in a source input image with each of the 100 largest faces detected in a target input image. Following are the fields related to this operation.

Field Required Supports Variables Description
(Source Image) Image Source :white_check_mark: :x: Whether the source image is provided as a reference to an S3 object or as base64-encoded image bytes
(Source Image) Bucket Name :white_check_mark: :white_check_mark: Name of the bucket containing the source image
(Only applicable if the Image Source is S3)
(Source Image) File Name :white_check_mark: :white_check_mark: S3 object key of the source image
(Only applicable if the Image Source is S3)
(Source Image) File Version :x: :white_check_mark: If versioning is enabled, S3 object version of the source image
(Only applicable if the Image Source is S3)
(Source Image) Content :white_check_mark: :white_check_mark: Blob of source image bytes up to 5 MBs
(Only applicable if the Image Source is Bytes)
(Target Image) Image Source :white_check_mark: :x: Whether the target image is provided as a reference to an S3 object or as base64-encoded image bytes
(Target Image) Bucket Name :white_check_mark: :white_check_mark: Name of the bucket containing the target image
(Only applicable if the Image Source is S3)
(Target Image) File Name :white_check_mark: :white_check_mark: S3 object key of the target image
(Only applicable if the Image Source is S3)
(Target Image) File Version :x: :white_check_mark: If versioning is enabled, S3 object version of the target image
(Only applicable if the Image Source is S3)
(Target Image) Content :white_check_mark: :white_check_mark: Blob of target image bytes up to 5 MBs
(Only applicable if the Image Source is Bytes)
Similarity Threshold :x: :white_check_mark: The minimum level of confidence in the face matches that a match must meet to be included in the result

Detect Faces

Detect Faces operation can be used to detect the 100 largest faces in a provided image. Following are the fields related to this operation.

Field Required Supports Variables Description
Image Source :white_check_mark: :x: Whether the image is provided as a reference to an S3 object or as base64-encoded image bytes
Bucket Name :white_check_mark: :white_check_mark: Name of the bucket containing the image
(Only applicable if the Image Source is S3)
File Name :white_check_mark: :white_check_mark: S3 object key of the image
(Only applicable if the Image Source is S3)
File Version :x: :white_check_mark: If versioning is enabled, S3 object version of the image
(Only applicable if the Image Source is S3)
Content :white_check_mark: :white_check_mark: Blob of image bytes up to 5 MBs
(Only applicable if the Image Source is Bytes)
Facial Attributes :x: :x: Which facial attributes to be returned in the result.

If a value is not selected or DEFAULT is selected, following subset of facial attributes will be returned.
BoundingBox, Confidence, Pose, Quality, and Landmarks

If ALL is selected, all facial attributes will be returned.

Detect Labels

Detect Labels operation can be used to detect instances of real-world entities within an image (JPEG or PNG) provided as input. This includes objects like flower, tree, and table; events like wedding, graduation, and birthday party; and concepts like landscape, evening, and nature. Following are the fields related to this operation.

Field Required Supports Variables Description
Image Source :white_check_mark: :x: Whether the image is provided as a reference to an S3 object or as base64-encoded image bytes
Bucket Name :white_check_mark: :white_check_mark: Name of the bucket containing the image
(Only applicable if the Image Source is S3)
File Name :white_check_mark: :white_check_mark: S3 object key of the image
(Only applicable if the Image Source is S3)
File Version :x: :white_check_mark: If versioning is enabled, S3 object version of the image
(Only applicable if the Image Source is S3)
Content :white_check_mark: :white_check_mark: Blob of image bytes up to 5 MBs
(Only applicable if the Image Source is Bytes)
Max Labels :x: :white_check_mark: Maximum number of labels to be return in the result. The specified number of highest confidence labels will be returned.
Min Confidence :x: :white_check_mark: The minimum level of confidence for the labels to be included in the result

Detect Text

Detect Text operation can be used to detect text in the input image and to convert it into machine-readable text. Following are the fields related to this operation.

Field Required Supports Variables Description
Image Source :white_check_mark: :x: Whether the image is provided as a reference to an S3 object or as base64-encoded image bytes
Bucket Name :white_check_mark: :white_check_mark: Name of the bucket containing the image
(Only applicable if the Image Source is S3)
File Name :white_check_mark: :white_check_mark: S3 object key of the image
(Only applicable if the Image Source is S3)
File Version :x: :white_check_mark: If versioning is enabled, S3 object version of the image
(Only applicable if the Image Source is S3)
Content :white_check_mark: :white_check_mark: Blob of image bytes up to 5 MBs
(Only applicable if the Image Source is Bytes)

Recognize Celebrities

Recognize Celebrities operation can be used to detect and recognize celebrities in the input image. Following are the fields related to this operation.

Field Required Supports Variables Description
Image Source :white_check_mark: :x: Whether the image is provided as a reference to an S3 object or as base64-encoded image bytes
Bucket Name :white_check_mark: :white_check_mark: Name of the bucket containing the image
(Only applicable if the Image Source is S3)
File Name :white_check_mark: :white_check_mark: S3 object key of the image
(Only applicable if the Image Source is S3)
File Version :x: :white_check_mark: If versioning is enabled, S3 object version of the image
(Only applicable if the Image Source is S3)
Content :white_check_mark: :white_check_mark: Blob of image bytes up to 5 MBs
(Only applicable if the Image Source is Bytes)

results matching ""

    No results matching ""