Generate photos based on over 1M protos processed by ourselves before

Question

We are running a huge team that process child photos for our customers, the team processes over 1M photos per year.

The process includes basic tuning of light, resize, apply some filters to make the skin looks better.

We want to use deep learning to complete the jobs as much as possible. Which means I want to choose one model and train that model using our existing data. And then use the trained model to generate photos by inputing the new unprocessed photos.

Is there existing model that I can make use of, or any papers have covered this scenario?

Any help would be appreciated, thanks!

score 1 · Answer 1 · answered Jan 04 '17 at 08:31

1

You could try something like this: https://arxiv.org/pdf/1412.7725.pdf. But with deep learning and your amount of training data you can problem get any big enough model to work well.

answered Jan 04 '17 at 08:31

Thomas Pinetz

6,948
2
27
46

You mean the amount is not enough? – Bin Chen Jan 04 '17 at 08:48
No i meant if you did this for several years you have several million images which is enough to train pretty much any model to perform well, given it has enough parameters to perform well in the first place. – Thomas Pinetz Jan 04 '17 at 08:49
Thanks. but I don't know how to start. is it possible we discuss thru email? I need some advice, thanks. – Bin Chen Jan 04 '17 at 10:56
Check out the getting started section in the documentation for deep learning. Choose your framework and programming language and then if you have any specific question you can ask again here. If you do not have any previous deep learning experience, i can recommend keras. – Thomas Pinetz Jan 04 '17 at 11:21

score 1 · Answer 2 · answered Jan 05 '17 at 19:58

Image generation is not what you should search for. Image generation means that an image is generated (almost) completely from nothing. You want to enhance an existing image.

Although I haven't read any papers about this scenario so far, searching for "image enhancement neural network" reveald several promising results:

A Survey on Image Enhancement Techniques: Classical Spatial Filter, Neural Network, Cellular Neural Network, and Fuzzy Filter: http://ieeexplore.ieee.org/document/4237993/
A new class of nonlinear filters for image enhancement: http://ieeexplore.ieee.org/document/150915/
An image enhancement technique combining sharpening and noise reduction: http://ieeexplore.ieee.org/document/1044761/

I guess you could do the following:

Create a CNN model. The only "special" thing of this model is that it does not have a fully connected layer as target, but another (3 channel) image. You have to adjust the error function to this. (Similar to semantic segmentation).

Generate photos based on over 1M protos processed by ourselves before

2 Answers2