Cases de use de Amazon Rekognition

In today's post we are going to talk about a recognition service that we already talked about in our post more generally , on this occasion we are going to delve into the use cases of Amazon Rekognition.

If you find this topic interesting, we invite you to download our free Ebook «How to migrate to Amazon Web Services?«

Labeling and recognition function in detail

Thanks to the function of smart tagging or automatic tagging, the user can quickly locate certain specific elements. This is especially useful when these elements must be located in huge quantities of files. And it should be noted that Amazon Rekognition works with both image files and video files.

We have 4 distinctive cases:

• Detecting inappropriate content ❌

• Online Identity Verification ✅

• Optimizing multimedia content analysis 📈

• Sending smart alerts from the connected home 🏠

1. Detection of inappropriate content

Rekognition helps us quickly and accurately identify unsafe or inappropriate content in images and videos based on your business's general standards or practices, 

Moderating this content allows you to automate and streamline image and video moderation workflows by Machine Learning (ML) without having prior experience in ML.

This tool allows you Process millions of images and videos efficiently and detect unwanted content with fully managed APIs and customizable moderation rules to be able to continue with a secure line for users and the company. Always with a service without minimum fees, licenses or initial commitments.

An example of detection that was applied during social distancing in the COVID-19 era, was the use of the Bounding Box from Rekognition to determine the coordinates of each of the people identified in the image and using that data, (using pixels they determined the distance in pixels, which separated each of the people in the queue through a reference value and classified it as non-compliance or not).

2. Online identity verification

According to AWS, in-person user identity verification is slow to scale and expensive, and poses significant pain points for users. Machine learning-based facial biometrics can enable online identity verification of users. 

Rekognition offers facial recognition and analysis capabilities that can be quickly included in the user onboarding and authentication workflow to verify the identity of users online. 

You don't need to be a machine learning expert, thanks to Rekognition you can onboard and authenticate users in seconds while detecting fraudulent or duplicate accounts. In this way we get Increase the number of users faster, reduce the number of frauds and lower costs user verification.

Identity verification therefore allows us to:

• Reduce fraud

• Minimize costes and general expenses

• Selfie validation with face detection

• Selfie vs. User ID comparison

• Detecting duplicate users

• Classification of identity documents

• Extraction of user data

3. Optimizing multimedia content analysis

Rekognition, within all its functionalities, allows you to automatically detect key video segments to reduce the time, effort and costs of video ad insertion, content operations and production. Also facilitates the automation of operational media analysis by providing fully managed, purpose-built APIs powered by ML.

These APIs make it easy to analyze large volumes of video stored in Amazon S3, detect shot changes, and obtain timecodes and timestamps.

Benefits:

• Streamlining operational tasks

• Obtaining accurate results

Features:

• Black frame detection

• Credit detection

• Color bar detection

• Detection of whiteboards and studio logos

4. Sending smart home alerts

With Amazon Rekognition you can send alerts when a desired object is detected in your live video streams, such as Create home automations, such as automatically turning on the light when a person is detected, also tand allows find learning solutions automatic at a low cost to process live video streams from cameras to detect people and objects in real time and provide actionable alerts to end users.

Benefits:

• Send relevant alerts in real time

• Improving “connected home” experiences

• Integration with Amazon Kinesis Video Streams

Amazon Rekognition in action

Here's a perfect example of some of the tagging capabilities mentioned above. During the AWS re:Invent event, it was shown how Rekognition is able to moderate the content that users can upload to various social media platforms. A great example of this is Pinterest. We're sharing the video with you.

At apser we have experts who will accompany you during your journey to the cloud and will be able to resolve any of your doubts, Contáctanos to study your project. We will create a roadmap to implement the solution that best suits your objectives.