Problem with recognize the real age

[roysG]

Hi, Thank you for your time for building this magic. But, there are leak in many images between the person and result i get on the user.

How can i improve the result?

[GilLevi]

Hi,

Thank you for your interest in out work.

I didn't understand your question. What do you mean by "leak in many images between the person and result I get on the user" ?

Best, Gil

[roysG]

Hi Gil, Lets say you check image of girl in age 17, but in the result you will get 0-2 or 60+.

I also have another question, maybe you can also assist me. I work with api of "deepdetecet" and when i am doing a lot of request(100 requests by second) i get a lot of errors "timeout". Maybe you have idea why does it happen or you have suggest any solution?

[GilLevi]

Hi,

1. Can you please upload the images so I can test them?
2. I never head of deepdetect. Sorry.

Best, Gil

[roysG]

Hi Gil, It happend me also in other images, but you can start with testing those images:

http://oi63.tinypic.com/2056tjd.jpg http://oi67.tinypic.com/2s6tudx.jpg http://oi67.tinypic.com/2dkxt0x.jpg

[GilLevi]

I see,

Did you crop the faces before running the network?

Best, Gil

[roysG]

Also tried both options

[GilLevi]

OK, the network expects cropped faces. Can you upload some examples of the cropped faces that you tried and what were the results you got? I'll try it with my code.

[roysG]

Hi Gil, I will looking for those images and update you.

Gil, i also want to install copy of caffe compiled version in amazon lambda.

Can you give me some tips how to do it?

[GilLevi]

Hi,

I'm not that familiar with AWS labmda, but I think you can find a lot of AMI which have Caffe installed. Hope that helps.

Best, Gil