rekognition index_faces : Unable to get object metadata from S3

Hi, I created a s3 bucket named: digiphoto-lumen-avventori-pdv-test (without special chars) in the us-east-1 region. I created a Role with: CloudWatchLogsFullAccess, AmazonS3FullAccess, AmazonDynamoDBFullAccess, AmazonRekognitionFullAccess I created a lambda function and a trigger for faces indexing.

I created the dynamodb table in same region (us-east-1) and the collection like this:

aws rekognition create-collection --collection-id avventori --region us-east-1

but when I uploading an image, the trigger raise the hateful exception: * An error occurred (InvalidS3ObjectException) when calling the IndexFaces operation: Unable to get object metadata from S3. Check object key, region and/or access permissions.*

Trying to detect faces from powershell CLI

aws rekognition detect-faces --region us-east-1 `
--image '{""S3Object"" : {""Bucket"" : ""digiphoto-lumen-avventori-pdv-test"", ""Name"" : ""image1.png"" }}'

the request's succeded (200 OK with correct response) Why python script fails ? What is the problem ? The key is correct, the region is correct the bucket is accessible with full permissions.

the python script starts with region:

s3 = boto3.client( 's3', region_name='us-east-1' )
rekognition = boto3.client( "rekognition", region_name='us-east-1' )
dynamodbTableName = "avventori"
dynamodb = boto3.resource( 'dynamodb', region_name='us-east-1' )
avventoriTable = dynamodb.Table( dynamodbTableName )

Cannot exit from this impasse. Can anyone give me a suggestion ? thanks