Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>Some method I had experience with them are </p> <ul> <li><a href="http://hal.inria.fr/inria-00439290" rel="nofollow">metric learning</a> for comparing faces</li> <li><a href="http://www.robots.ox.ac.uk/~vgg/research/nface/" rel="nofollow">naming video characters</a>: they use SIFT descriptors computed at specific feducial points on each face. Their code worked quite well for me in the past.</li> </ul> <p>A dataset and benchmark that is dedicated for this task is <a href="http://vis-www.cs.umass.edu/lfw/" rel="nofollow">labeled faces in the wild</a>. You can find there references to working methods for comparing faces after detection.</p> <p><strong>UPDATE:</strong><br> I have a description of an experiment on face clustering: unsupervised face identification. The experiment is described in <a href="http://arxiv.org/pdf/1210.7362v2.pdf" rel="nofollow">Section 4.4 of my thesis</a>.<br> The basic flow is as follows</p> <ol> <li><p><strong>Metric learning:</strong> how to determine if two faces are of the same person or not.<br> This part is supervised, in the sense that it requires as input face images labeled with the identity of the person who appears in each photo.</p> <p>a. Detect fiducial points (eyes, corner of mouth, nose).<br> You may use <a href="http://www.robots.ox.ac.uk/~vgg/research/nface/" rel="nofollow">this code</a>, or more recent versions such as <a href="http://www.ics.uci.edu/~xzhu/face/" rel="nofollow">this one</a>.</p> <p>b. Extract SIFT descriptors at the detected fiducial points. </p> <p>c. Construct a "face descriptor": each face is described using a <strong>single</strong> vector.<br> This vector is a concatenation of the <strong><code>sqrt</code></strong> of all the SIFT descriptors.</p> <p>d. Use the method described <a href="http://hal.inria.fr/inria-00439290" rel="nofollow">here</a> to learn a mahalanobis distance between faces of different persons.</p></li> <li><p><strong>Unsupervised face identification:</strong> Once a metric was learned, you may use new photos of <strong>new</strong> people (these people need not be part of the training set, you may use photos of <strong>unseen-before</strong> people!). </p> <p>a. Repeat stages a-c to construct the same "face descriptor" vector for each input face.</p> <p>b. Compare the descriptor vectors using the learned mahalanobis distance.</p></li> </ol>
    singulars
    1. This table or related slice is empty.
    1. POFace identification with opencv
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USShai
    1. USShai
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    3. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTBountyClose
    1. COA thesis-based answer! So you've done some research in this field. Thanks for sharing your findings.
      singulars
      1. PO
      1. USikel
    2. CO@IwanKelaiah thanks. I hope you'll find it useful. I have some relevant Matlab code on [my web page](http://www.wisdom.weizmann.ac.il/~bagon/matlab.html).
      singulars
      1. PO
      1. USShai
    3. COMost useful. Thank you.
      singulars
      1. PO
      1. USikel
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload