[Report] AIM305-R [REPEAT] Automatically Extract Metadata Using Computer Vision & Language AI Services #reinvent

2018.12.09

この記事は公開されてから1年以上経過しています。情報が古い可能性がありますので、ご注意ください。

I participated in the AIM305-R [REPEAT] Automatically Extract Metadata Using Computer Vision & Language AI Services of the builders session.

The builders session is a group session with AWS Solution Architect and five participants.

Builders Sessions

Overview

Here is the session summary from the Event Catalog.

Customers are using automatic metadata extraction to fuel new insights and provide innovative services to their customers. In this session, we walk through the basic architecture patterns for implementing automatic metadata extraction using Amazon Rekognition, Amazon Transcribe, and Amazon Comprehend. We also share how to get started with the pre-configured AWS Media Analysis Solution.

Solution

画像

Deployment

We will present AWS's solution architect, and will build it according to the following documents:

Step 1. Launch the Stack

We will use CloudFormation template (media-analysis-deploy.template) prepared in advance.

Enter Stack name and Email address.

  • Specify Details
    • Stack name:
  • Parameters
    • Email address:

画像

The email address will be created in the Amazon Cognito Identity Pool. Steps to log into the web site will be sent to the email address.

Check the following and click Create.

  • Capabilities
    • I acknowledge that AWS CloudFormation might create IAM resources with custom names.
    • I acknowledge that AWS CloudFormation might require the following capability: CAPABILITY_AUTO_EXPAND

画像

We will access the web site from the email link.

画像

Step 2. Upload a File for Processing

We will attempt to sign up and upload images and videos.

画像 画像 画像

We can also check the progress in the AWS Step Functions console from View progress in your AWS Console.

画像 画像

Step 3. View Extracted Metadata

After extracting the metadata, we can check the extracted metadata by clicking View results.

  • Labels

画像

  • Facial Attributes

画像

Conclusion

The topic Artificial Intelligence & Machine Learning is popular nowadays, and from this session, I learnt that builders session can actually be constructed and we can grasp the image of the solution. Since there were only a few people, we were able to ask questions and hold open discussions.

Reference