Guide to analysis operators - AWS Content Analysis

Guide to analysis operators

When uploading videos, users can select which of the following services to include in the analysis workflow that starts after upload completes:

  • Amazon Rekognition to:

    • Detect objects in videos

    • Detect celebrities in videos

    • Detect faces in videos

    • Detect potentially unsafe or inappropriate content in videos

  • Amazon Transcribe to convert speech to text from video audio tracks

  • Amazon Translate to convert text from one language to another

  • Amazon Comprehend to:

    • Identify entities in text

    • Identify key phrases in text

By default, all analysis operators are activated, as shown in Figure 6.

Use Amazon Rekognition to detect objects and activities in videos

Use the following procedure to detect objects and activities in videos using Amazon Rekognition.

  1. Sign in to the AWS Content Analysis web application.

  2. Choose Upload.

  3. From the Upload Videos page, drag and drop one or more video files into the upload box.

  4. Choose Configure Workflow and verify that the Amazon Rekognition features are activated including object detection and other detection options in the workflow settings dialog.

  5. Choose Upload and Run Workflow. Wait for the workflow status to show as complete before continuing to the next step.

  6. Choose Collection.

  7. Locate the media file you want to analyze and under the Actions column choose Analyze.

  8. From the ML Vision tab, explore the Amazon Rekognition metadata for your content.

Use Amazon Transcribe to convert speech to text from audio and video assets

Use the following procedure to convert speech to text from video assets using Amazon Transcribe.

  1. Sign in to the AWS Content Analysis web application.

  2. Choose Upload.

  3. From the Upload Videos page, drag and drop one or more media files into the upload box.

  4. Choose Configure Workflow and verify that the Transcribe option is selected.

  5. In the Source Language drop-down list, select the language you want.

  6. Choose Upload and Run Workflow. Wait for the workflow status to show as complete before continuing to the next step.

  7. Choose Collection.

  8. Locate the media file you want to analyze and under the Actions column, choose Analyze.

  9. From the Speech Recognition tab, explore the Amazon Transcribe metadata for your content.

Use Amazon Comprehend to identify entities in text

Use the following procedure identify entities in text using Amazon Comprehend.

  1. Sign in to the AWS Content Analysis web application.

  2. Choose Upload.

  3. From the Upload Videos page, drag and drop one or more media files into the upload box.

  4. Choose Configure Workflow and verify that the Transcribe option is selected.

  5. In the Source Language drop-down list, select the language you want.

  6. In the Text Operators box, verify that the Comprehend Entities option is selected.

  7. Choose Upload and Run Workflow. Wait for the workflow status to show complete before continuing to the next step.

  8. Choose Collection.

  9. Locate the media file you want to analyze and under the Actions column, choose Analyze.

  10. From the Speech Recognition tab, explore the Amazon Comprehend metadata for your content.