wongcyrus / HowWhoFeelInVideo

MIT License
6 stars 4 forks source link

The process "Create Emoji Video" failed to run #2

Open Daniel-1234567 opened 6 years ago

Daniel-1234567 commented 6 years ago

Hi, I encountered an issue. I can not solve it though I have tried many many times. I think there may be something wrong.

The major problem is that the lambda function (Cascades Face Detection or Process Images, I think) cannot convert all the .jpg images to .png images. In the Step Functions Dashboard, the Cascades Face Detection lambda seems to be normal.

But, look from CloudWatch console, when the ProcessImage lambda process the images, some images can be processed in a normal way, while some image processes log error: The collection id: student does not exist (Service: AmazonRekognition; Status Code: 400; Error Code: ResourceNotFoundException;

And, in the CreateEmojiVideo's log, the error log is: 2018-04-12T03:58:15.928Z b9e2d831-3e05-11e8-b5dd-87544478c7b0 { "message": "The specified key does not exist.", "code": "NoSuchKey", "region": null, "time": "2018-04-12T03:58:15.926Z", "requestId": "AE061CC0E83D1C3A", "extendedRequestId": "zKwitnGQ7WpBjRopI3IwjuR88MX7HN49r4TUFcuomdv6kR4f8lkAd5wJwXdtle+eZXsVgktVkyY=", "statusCode": 404, "retryable": false, "retryDelay": 74.93055357424083 }

Could you please help me solve this problem? Thank you!

capture capture1

Daniel-1234567 commented 6 years ago

And I have enabled the X-ray. And I have added face collections step by step.

wongcyrus commented 6 years ago

Are you deploy the system in the region with Amazon Rekognition? Also, have you added some images to the collection?

Daniel-1234567 commented 6 years ago

Thanks for your response. Yes, I deployed the stack in N.Virginia (us-east-1). And I have added about 6 images to the collection student. list-faces

Daniel-1234567 commented 6 years ago

I do not know how to solve the problem in the prerequisite that I do not modify your lambda code or scala code. Is it a good idea that I modify your lambda code or scala code? Thanks, sir.