Alpkant / Thermal-to-Visible-Face-Recognition-Using-Deep-Autoencoders

Official repository of our BIOSIG19 paper "Thermal to Visible Face Recognition Using Deep Autoencoders"
MIT License
35 stars 6 forks source link

how to get training data? #3

Open MujoNoThuki opened 3 years ago

MujoNoThuki commented 3 years ago

I'm new to python, so it may be a silly question, but let me ask you a question. There is a file such as "01.xml" in "Carl_annotations". I understand that the name of each thermal image file and data such as the coordinates of the facial parts in the image are written there. But how do I get this image file?

Alpkant commented 3 years ago

Hi, I am not sure whether I understood your question right but you can download the dataset from http://splab.cz/en/download/databaze/carl-database. However, you should fill the license agreement form and get access.

MujoNoThuki commented 3 years ago

Thank you for your reply while you are busy. I have already sent an application to the site the other day and received a link to download the dataset.

However, while the name of the image file described in "01.xml" of "Carl_annotations" is "01 / session_4 / thermal_bmp / 01S4BNA5.bmp", the actually downloaded data set I couldn't find any image files with such names and folders with names like "01" and "session_4".

Alpkant commented 3 years ago

It looks like the authors of the dataset changed the folder structure and naming convention. While I was using the dataset there was a directory for each subject such as "01", "02". Within that folder, there were session folders. I don't know the new naming convention but I will try to figure out and update the repo according to it.

MujoNoThuki commented 3 years ago

Oh, I see.