about the score of Attribute
Opened this issue · 5 comments
houxiaoxia1 commented
Hi, sir, what's the meaning of negative value in imgList*.txt? such as imgListTrainRegression_ColorHarmony.txt, farm1_429_20164497916_612b5cce64_b.jpg -0.400. Thank you.
houxiaoxia1 commented
In the paper, you choose image pairs within rater, could you share the file which has the scores from different raters? Thank you.
aimerykong commented
hi, negative values means the attribute has a negatively effect on the
image aesthetics. The annotation for each rater is in {-1,0,1}; given
annotation of five raters, we average their annotation, therefore, there
are values inbetween [-1,1].
…On Tue, Apr 3, 2018 at 5:14 AM, houxiaoxia1 ***@***.***> wrote:
Hi, sir, what's the meaning of negative value in imgList*.txt? such as
imgListTrainRegression_ColorHarmony.txt, farm1_429_20164497916_612b5cce64_b.jpg
-0.400. Thank you.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#16>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AGKZJOl8-b_tIJQWKKOQWD7b_trmKiFaks5tk2eMgaJpZM4TE9gM>
.
aimerykong commented
In the following link, you can find read the ".mat" file for all the
annotations, including raters' ID. Then you can generate training pairs in
your way.
https://drive.google.com/drive/folders/0BxeylfSgpk1MOVduWGxyVlJFUHM
…On Tue, Apr 3, 2018 at 5:22 AM, houxiaoxia1 ***@***.***> wrote:
In the paper, you choose image pairs within rater, could you share the
file which has the scores from different raters? Thank you.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#16 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AGKZJBsw-rjMYWO1b-Y-ixEGtr91r5Ehks5tk2mHgaJpZM4TE9gM>
.
houxiaoxia1 commented
Thank you!
houxiaoxia1 commented
when I trained a resnet to predict the score of a picture, I only got rank correlation statistic 0.56 rather than 0.6+. Could you tell me how to choose image pairs? I choose them with a interval of 0.15, how about you?