# TUGraz-TUT Face Verification Challenge — Data

来源：https://www.kaggle.com/competitions/face-verification2/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

<h2>File descriptions</h2>
<ul>
<li><strong>train.zip</strong> - the training set containing several images per person. </li>
<li><strong>test.zip </strong>- images used for your verification submission (see pairs.csv)</li>
<li><strong>pairs.zip</strong> - a list of image pairs for your submission. </li>
</ul>
<p>The training database consists of a set of folders, where each folder contains pictures of a single person. You are free to sample pairs of same / different persons and learn to distinguish between the two cases. Also any other training data is allowed (but not mandatory) for this purpose (e.g. <a href="http://vis-www.cs.umass.edu/lfw/">LFW </a>or <a href="http://www.cs.tau.ac.il/~wolf/ytfaces/">YTF </a>datasets). The training database also contains the following readily calculated data for each image:</p>
<ul>
<li><strong>&lt;image_name&gt;_bbox.csv</strong> - the coordinates of the bounding box around the face. Contains four numbers: <em>x, y, width, height</em> with (<em>x,y</em>) the upper left corner.</li>
<li><strong>&lt;image_name&gt;_deepfeatures.csv</strong> - Deep neural network features computed at the last hidden layer of <em>AlexNet.</em></li>
<li><strong>&lt;image_name&gt;_landmarks_small.csv</strong> - Five landmark coordinates extracted from the image: <em>left eye, right eye, nose, left corner of mouth, right corner of mouth.</em></li>
<li>
<p><strong>&lt;image_name&gt;_landmarks_big.csv</strong> - 68 landmark coordinates extracted from the image.</p>
</li>
</ul>
<p>The test database consists of images similar to those in the training database. However, the identities are not given. Instead, a list of image pairs (the third file) is provided, and competitors are required to determine whether each pair is a match (1) or non-match (0).</p>
<p><strong>Note: </strong>The data was re-uploaded on October 28th, because the deep features were incorrectly computed. Please reload train.zip and test.zip.</p>