If i have a dataset consisting by a list of images each associated with a series of features; there is a model that, once trained, generates new images upon entering a new list of features?
Asked
Active
Viewed 527 times
-2
-
what do you mean by features? – deadcode Apr 05 '19 at 11:30
-
imagine you have a series of brain MRI images and their EEGs. I would like a network that inserts a new EEG to generate a possible MRI image – Francesco Torre Apr 05 '19 at 12:04
1 Answers
2
I think you are looking for GAN(Generative Adversarial Networks) which is proposed in this paper.
GAN are the type of algorithm which contains two different model so that one model named Discriminator tries to learn to determine if it's input data comes from the data set or not and the other one named Generator tries to learn how to generate data so that the Discriminator wrongly recognize that it comes from the data set.
You can find more details from the following links:
generative adversarial network (GAN)
Generative Adversarial Networks (GANs): Engine and Applications
pouyan
- 3,445
- 4
- 26
- 44
-
if I understand correctly GAN generates new random images, I would like to generate them in relation to the new data entered – Francesco Torre Apr 05 '19 at 11:53
-
You can play with the model according to your needs and customize it. In your case I think one possible solution is to use the entered data as Latent space or even you can use it as noise data. – pouyan Apr 05 '19 at 11:58
-
It is generally a nice move to attribute other people's images when using them: [Source](https://www.linkedin.com/pulse/gans-one-hottest-topics-machine-learning-al-gharakhanian) – m00am Apr 05 '19 at 12:20
-
@m00am, You are right, But I used the image from [this link](https://blog.statsbot.co/generative-adversarial-networks-gans-engine-and-applications-f96291965b47) and I mentioned it in my answer(second link). – pouyan Apr 05 '19 at 12:24