In this tutorial, you will learn how to use the video analysis features in Amazon Rekognition Video using the AWS Console. Amazon Rekognition Video is a deep learning powered video analysis service that detects activities and recognizes objects, celebrities, and inappropriate content.

As a developer, analyzing video is a challenge you will face if you are developing a video cataloging system or creating an application to provide sentiment analysis. This challenge can be solved by building your own machine learning model, however this option is time-intensive, expensive, and requires machine learning expertise.

Amazon Rekognition Video provides an easy-to-use API that offers real-time analysis of streaming video and facial analysis. This fully-managed, API-driven service enables developers to easily add visual analysis to existing applications.

In this tutorial, you will use Amazon Rekognition Video to analyze a 30-second clip of an Ultimate Frisbee game. Analyzing video automatically extracts rich metadata, which can be used to create a searchable video library, perform content moderation, or provide personalized VIP experiences.

This tutorial is a demo of the functionality that is available when using the AWS CLI or the Rekognition API. For production or proof of concept implementations, we recommend using these programmatic interfaces rather than the Amazon Rekognition Console.

a.  Open the AWS Management Console, so you can keep this step-by-step guide open. When the screen loads, enter your user name and password to get started. Then type Rekognition in the search bar and select Rekognition to open the service console.

a) To start, select Video analysis in the panel navigation on the left. This feature allows you to analyze video and receive a JSON response.

b) The pre-uploaded video of Jeff Bezos and Werner Vogels demonstrates how Rekognition Video can track people, detects activities, and recognize objects, celebrities, and inappropriate content.

c) First, under People, click on the Werner Vogels icon. Notice that the video on the right now shows when Werner was in the clip.

d) Next, notice the object and activities that were detected in this short clip. You can click on Beard and see the exact times Werner’s beard was detected or click on Furniture to see when the chairs were detected.

e) Also, you can see that no moderated labels were detected. This feature enables you to determine and filter inappropriate content. For example, you might want to filter images that contain nudity, but not images containing suggestive content.

a) Find a video you would like to use in this tutorial. The screenshots shown below are a combination of the sample video of Jeff Bezos and Werner Vogels, as well as an example of a soccer video uploaded to the console.

Note that video files can’t be larger than 60 mb or longer than 60 seconds to demo for free in the console.

b) Click the down arrow under Choose a sample or upload your own, click Your own video, and select the video footage from your computer.

After 40-50 seconds, the video will be analyzed and the results will be visible in the console.

c) In this example, notice that Rekognition detected 12 people from the clip. For example, click on Show more under People and select the referee in the orange and black striped shirt. You can see in the video analysis on the right the times when this particular referee was detected in the video.

d) Click on Objects and activities. Notice that there are 20 objects and activities that were automatically tagged. For example, click on Automobile and see the scenes in the video when an automobile was detected.

e) Next, click on Team Sport and notice that each flagged clip on the video analysis on the right will contain at least one the labels we’ve selected throughout this step: the referee under People, an automobile as the Object, and Team Sport as the Activity.

As a developer, the automatic metadata from video footage can be used in applications to improve marketing and advertising efforts, create a searchable video library, or set up sports tracking to provide rich analytics for sport games.

a) Click on Services and select S3.

b) The video you uploaded was automatically saved to a S3 bucket and you will need to delete it to ensure you don’t incur a cost. Scroll through your S3 buckets and find the bucket that starts with rekognition-video-console-demo. Click on this bucket and select all media and then right click and select Delete.