Automate Your Media Supply Chain with Machine Learning
- 1. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
- 2. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. AmazonRekognition DeepLearning-basedimageandvideoanalysis
- 3. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. AmazonRekognitionfeatures Faces Celebrities Objects Moderation ScenesActivities Paths Text
- 4. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. TextinImage
- 5. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Languageservices
- 6. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Serverless Media Supply Chain Amazon Rekognition Amazon Transcribe Amazon Comprehend Amazon S3 AWS Step Functions AWS Lambda Amazon Elasticsearch Service image video audiotext
- 7. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. How are customers and partners using this today?
- 8. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Automated subtitle generation and translation
- 9. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Tag, filter, and redact inappropriate content Person 99.2% Gun 84.6% Handgun 73.5% redact faces detect weapons Drink 96.4% Alcohol 80.1% Wine 69.9% tag culturally sensitive content
- 10. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
- 11. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
- 12. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
- 13. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
- 14. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
- 15. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
- 16. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Editor's Notes
- #7: <unhide this slide when technical detail is needed for the audience> Media asset tagging is an approach that enables moderation and compliance requirements to be extended across image, video, speech, and text in a seamless pipeline that scales to accommodate millions of media assets.
- #10: In addition to face redaction, you may need to index video, image, and even textual content to detect culturally sensitive themes, like alcohol and smoking. Videos may earn a specific rating if weapons are present or shooting occurs, both of which can be detected with high accuracy using AWS AI services.