Amazon Rekognition is a computer vision service. It makes use of Deep Learning models to classify images.It is majorly used for video and image analysis. It can detect objects, scenes and actions.

It takes inputs in the form of images, videos, and photo collections. To work on this no prior knowledge on ML or DL is required.It generates high accuracy facial search and analysis.

In Rekognition, we pay for the input analyzed and the metadata of the analysis stored.

Features:

1.Label Detection:Rekognition can detect objects, activities, scenes and landmarks.

1. Image Properties: Rekognition can identify dominant colors, image brightness,sharpness,contrast and other image properties.

2. Face Search: It allows for searching faces using a collection of input faces and matching them.

3. Facial Analysis: It detects faces in an image and analyze characters like smile, open eyes,etc..

1. Text Detection: Rekognition detects text in images and videos and converts them to machine readable format.

2. Content Moderation: Rekognition identifies and moderates inappropriate content, ensuring safer user experience.

Benefits and Use Cases:

• Enhanced Authentication
• Image Search
• Autonomous Driving Cars
• Real time Analysis
• Scalablity and Integration
• Simple, Easy to Use
• Analyse and Filter Image Properties
• Low Cost.

How It Works?

Rekognition provides 2 API sets for Analysis for image and video inputs, viz:

-Amazon Rekognition Image
-Amazon Rekognition Video
**
Amazon Rekognition Image:**

Rekognition Image offers a set of capablities for image analysis. It can detect objects and concepts in an image.It used in recognizing celebrities,detects text (text in any language),identifies explicit content. It is majorly used in analyzing face attributes like emotion. It has an ability to detect objects like masks, helmets, gloves or safety vests.This ablity is known as Detecting the Presecnce of Personel Protective Equipment(PPE).These features enables various applications like image organization, moderating contents,etc..

Amazon Rekognition Video:

Amazon Rekognition Video provides a range of features for analyzing video content. It can track people and objects as they move through different frames of a video, recognize various objects and even celebrities, and search through stored or streaming videos to find specific individuals. This technology can also analyze faces within the video, identifying attributes like age and emotions, and flagging explicit or inappropriate content. Also, it can organize and sort analysis results based on timestamps and segments, detect people, pets, and packages in real-time streaming video.These Features enable a various applications like analyzing video content, organizing video libraries,moderating contents, etc..

Rekognition Liveliness:

Rekognition Liveliness resolves spoofing attacks on Face Recognition system.Spoofing attacks refers to the activity where users try to decieve the system by presenting fake images or facial features.

Rekognition Liveliness aims to verify the persons presence in the location. This is detected by face movement, blinking, expressions, etc..