Basic facts on Amazon Rekognition

shrimali senevirathna
3 min readDec 10, 2020


Rekognition is an AWS offering in Artificial Intelligence. It is a cloud-based Software as a service. It allows to easily analyze images. It recognizes faces , landscapes and even moods of a person. It also allows you to search and compare faces. Here are some benefits you can gain by using Rekognition Service.

Easy to Use

It does not require a massive technological knowledge because it is simply uploading images to the S 3 bucket and all the other behind the scenes are handled by the engine as it provides image recognition at the push of a button. The best part is the available API s which are used to automate the processes.

Low Cost

Rekognition currently offers free tier access, up to 5,000 API calls per month. After that, you will be charged per image call according to the latest pricings.


It provides the facility to send thousands of images in a small time to check it faster.

Below are the main services provided by the AWS rekognition.

  1. Object Scene detection
  2. Facial Analysis
  3. Face comparison
  4. Facial rekognition
  5. Confidence scores on processed images

From object Scene detection is accurately determines what is in an image. For example, a photo of a person on a beach may contain labels such as Swimwear (objects), Beach (scene).So we can say that it simply recognizes and labels Objects and Scenes (landscapes).

In summary facial recognition, it compares the features to the provided image and finds the perfect match and in face comparison , it compares and differentiates the given image to the set of images to find out the most similar one. In that situation it provides and confidence score to get a conclusion whether the accuracy of the process is high or low. And also facial analysis can detect the mood of a person’s image.

Here is a common and a simple use case of Rekognition.

Image moderation


Image moderation is a process that provides a space for the moderators to ensure that inappropriate content is not posted. The process is clearly displayed in the above image. When uploading an image the S3 bucket object upload triggers Lambda function. Lambda calls rekognition which examines for the content. If there were no any inappropriate content detected the image is posted and can be viewed. But if any inappropriate content is detected ,the post is rejected or else it could be posted as a manual view.

It is amazing how the technologies are becoming more powerful. Not only Rekognition , there are many platforms that we can use for our future researches and for our developments in organizations and in services.