2
$\begingroup$

I have labeled faces images (label is the age -continuous value-) dataset and I want to construct a Convolutional Neural Network model to predict the age of a person.

I have the following questions.

  • How the label i.e. the age must be encoded ?
  • Which activation function to use ?
$\endgroup$
1
  • $\begingroup$ yes, in fact i have another outputs (ethnicity and gender) and its okay for them, but i still have a problem with age, i don't know if i have to do some modifications in age encoding or in the neural net layers! $\endgroup$ – Skander Hamdi Feb 27 '19 at 9:47
0
$\begingroup$

Convolutional networks can be used for regression tasks too. The difference corresponds to the output layers of the dense networks. In classification tasks you use sigmoid or softmax depending on your task. In regression tasks you can simply use linear activation function as the non-linearity of the last layers. Consequently, if you have the ages as the outputs of your network and if they are in a same scale as the inputs you can simply use them as they are right now.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.