Detecting anomalies in an image

To detect anomalies in an image with a trained Amazon Lookout for Vision model, you call the DetectAnomalies operation. The result from DetectAnomalies includes a boolean prediction that the image contains one or more anomalies and a confidence value for the prediction.

Images analyzed with DetectAnomalies must have the same dimensions as the images used to train the model.

Note

Before calling DetectAnomalies, you must first start your model with the StartModel operation. For more information, see Starting your Amazon Lookout for Vision model. You are charged for the amount of time, in minutes, that a model runs and for the number of anomaly detection units that your model uses. If you are not using a model, use the StopModel operation to stop your model. For more information, see Stopping your Amazon Lookout for Vision model.

To call DetectAnomalies, specify the following:

• Project – The name of the project that contains the model that you want to use.

• ModelVersion – The version of the model that you want to use.

• ContentType – The type of image that you want analyze. Valid values are image/png (PNG format images) and image/jpeg (JPG format images).

• Body – The unencoded binary bytes that represent the image.

The response from DetectAnomalies contains the following:

• IsAnomalous– A boolean indicator that the image contains one or more anomalies.

• Confidence– The confidence that Amazon Lookout for Vision has in the accuracy of the anomaly prediction (IsAnomalous). Confidence is a floating point value between 0 and 1. A higher value indicates a higher confidence.

• Source – Information about the image passed to DetectAnomalies.

{
    "DetectAnomalyResult": {
        "Source": {
            "Type": "direct"
        },
        "IsAnomalous": true,
        "Confidence": 0.9996867775917053
    }
}

If you're finding the confidence values returned by DetectAnomalies are too low, consider retraining the model.

To detect anomalies in an image (API)

1. If you haven't already done so, do the following:

   1. Create or update an IAM user. For more information, see Step 2: Create an IAM administrator user and group.

   2. Install and configure the AWS CLI and the AWS SDKs. For more information, see Step 5: Set up the AWS CLI and AWS SDKs.

   3. Train your model.

   4. Start your model.

2. Ensure the IAM user calling DetectAnomalies has access to the model version that you want to use.

3. Use the following examples to call the DetectAnomalies operation.

   Tip

   If you're using the Getting Started images, use an image from the extra_images folder.

   AWS CLI

   This AWS CLI command displays the JSON output for the DetectAnomalies CLI operation. Change the values of the following input parameters:

   • project name with the name of the project that you want to use.

   • model version with the version of the model that you want to use.

   • content type with the type of the image that you want to use. Valid values are image/png (PNG format images) and image/jpeg (JPG format images).

   • file name with the path and file name of the image that you want to use. Ensure the file type matches the value of content-type.

   aws lookoutvision detect-anomalies  --project-name project name\
       --model-version model version\
       --content-type content type\
       --body file name

   Python

   The following example code detects anomalies in an image that you supply. In the function main, change the following values:

   • project_name with the name of the project that contains the model version that you want to use.

   • model_version with the version of the model that you want to use.

   • photo with the name and file path of a JPG format image file that you want to use.

   • (Optional) If you want to pass a PNG format file, change the detect_anomalies input parameter ContentType='image/jpeg' to ContentType='image/png'.

   # Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.
   # SPDX-License-Identifier: Apache-2.0
   
   import boto3
   
   from botocore.exceptions import ClientError
   
   def detect_anomalies(project_name, model_version, photo):
       """
       Detects anomalies in an image (jpg/png) by using your Amazon Lookout for Vision model.
       param: project_name: - the name of the project that contains the model that you want to use.
       param: model_version: - the version of the model that you want to use.
       param: photo: - the path and name of the image in which you want to detect anomalies.
       """
   
       try:
   
           client = boto3.client("lookoutvision")
   
           # Call detect_anomalies
   
           with open(photo, "rb") as image:
               response = client.detect_anomalies(
                   ProjectName=project_name,
                   ContentType="image/jpeg",  # or image/png for png format input image.
                   Body=image.read(),
                   ModelVersion=model_version,
               )
           print("Anomalous?: " + str(response["DetectAnomalyResult"]["IsAnomalous"]))
           print("Confidence: " + str(response["DetectAnomalyResult"]["Confidence"]))
   
       except ClientError as err:
           print("Service error: " + format(err))
           raise
   
   
   def main():
   
       project_name = "my-project"  # Change to your project name.
       photo = "image.jpg"  # Chang to the path and name of the image (jpeg/png) that you want to use.
       model_version = "1"  # Change to the version of your model.
   
       detect_anomalies(project_name, model_version, photo)
   
   
   if __name__ == "__main__":
       main()
   .

4. If you aren't planning to continue using your model, stop your model.