Detect Image Labels with AWS Rekognition

$BucketName = 'trevorrekognition'
$FileName = 'kitchen.jpg'

New-S3Bucket -BucketName $BucketName
Write-S3Object -BucketName $BucketName -File $FileName
$REKResult = Find-REKLabel -Region us-east-1 -ImageBucket $BucketName -ImageName $FileName

$REKResult.Labels

After running the script above, you should have results printed in your PowerShell host that look something similar to the following:

RESULTS:

Confidence Name
---------- ----
86.87605   Indoors
86.87605   Interior Design
86.87605   Room
77.4853    Kitchen
77.25354   Housing
77.25354   Loft
66.77325   Appliance
66.77325   Oven

Using the AWS PowerShell module in conjunction with the AWS Rekognition service, you can detect labels in an image, such as identifying objects in a room, attributes about photos you took, and the corresponding confidence level that AWS Rekognition has for each of those attributes.

The Find-REKLabel command is the one that enables you to invoke a search for these attributes / labels. While you can provide image content as a byte array during the API call, a better method is to upload your image files to an AWS S3 Bucket, and then point the Rekognition service over to the S3 Objects that you want to analyze. The example above shows how to accomplish this.