Exercise 4 - using validation data in train

[ikvision]

@vedaldi Thank you for the detailed tutorial, I learned a lot by playing with it.
In exercise 4 it seems the imageMean is calculated on all the data (train+validation). In realistic scenario the validation set is given after the training phase.
What do you think about the following change:
exercise4.m on line 45:
imageMean = mean(imdb.images.data(:)) ;
should maybe replaced by
imageMean = mean(imdb.images.data(imdb.images.set==1)) ;

[vedaldi]

Hi, yes, this is probably a good suggestion. However, you would need to modify your solution a little to account for the fact that data has H x W x 3 x N dimensions.

On 26 Apr 2016, at 02:53, ik_vision [email protected] wrote:

@vedaldi https://github.com/vedaldi Thank you for the detailed tutorial, I learned a lot by playing with it.
In exercise 4 it seems the imageMean is calculated on all the data (train+validation). In realistic scenario the validation set is given after the training phase.
What do you think about the following change:
exercise4.m on line 45:
imageMean = mean(imdb.images.data(:)) ;
%should maybe replaced by
imageMean = mean(imdb.images.data( imdb.images.set==1)) ;

—
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub #14

[ikvision]

Thank you for the quick reply, I fixed my solution to:
train_data=imdb.images.data(:,:,imdb.images.set==1);
imageMean =mean(train_data(:))
It made Less than a half a percent diff in the imageMean value