<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1 plus MathML 2.0 plus SVG 1.1//EN" "http://www.w3.org/2002/04/xhtml-math-svg/xhtml-math-svg.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
  <head>
    <meta http-equiv="Content-Type" content="application/xhtml+xml; charset=utf-8"/>
    <title>Project-Team:WILLOW</title>
    <link rel="stylesheet" href="../static/css/raweb.css" type="text/css"/>
    <meta name="description" content="New Software and Platforms - NetVLAD: CNN architecture for weakly supervised place recognition"/>
    <meta name="dc.title" content="New Software and Platforms - NetVLAD: CNN architecture for weakly supervised place recognition"/>
    <meta name="dc.subject" content=""/>
    <meta name="dc.publisher" content="INRIA"/>
    <meta name="dc.date" content="(SCHEME=ISO8601) 2016-01"/>
    <meta name="dc.type" content="Report"/>
    <meta name="dc.language" content="(SCHEME=ISO639-1) en"/>
    <meta name="projet" content="WILLOW"/>
    <script type="text/javascript" src="https://raweb.inria.fr/rapportsactivite/RA2016/static/MathJax/MathJax.js?config=TeX-MML-AM_CHTML">
      <!--MathJax-->
    </script>
  </head>
  <body>
    <div class="tdmdiv">
      <div class="logo">
        <a href="http://www.inria.fr">
          <img style="align:bottom; border:none" src="../static/img/icons/logo_INRIA-coul.jpg" alt="Inria"/>
        </a>
      </div>
      <div class="TdmEntry">
        <div class="tdmentete">
          <a href="uid0.html">Project-Team Willow</a>
        </div>
        <span>
          <a href="uid1.html">Members</a>
        </span>
      </div>
      <div class="TdmEntry">Overall Objectives<ul><li><a href="./uid3.html">Statement</a></li></ul></div>
      <div class="TdmEntry">Research Program<ul><li><a href="uid5.html&#10;&#9;&#9;  ">3D object and scene modeling, analysis,
and retrieval</a></li><li><a href="uid7.html&#10;&#9;&#9;  ">Category-level object and scene recognition</a></li><li><a href="uid8.html&#10;&#9;&#9;  ">Image restoration, manipulation and enhancement</a></li><li><a href="uid9.html&#10;&#9;&#9;  ">Human activity capture and classification</a></li></ul></div>
      <div class="TdmEntry">Application Domains<ul><li><a href="uid13.html&#10;&#9;&#9;  ">Introduction</a></li><li><a href="uid14.html&#10;&#9;&#9;  ">Quantitative image analysis in science and humanities</a></li><li><a href="uid15.html&#10;&#9;&#9;  ">Video Annotation, Interpretation, and Retrieval</a></li></ul></div>
      <div class="TdmEntry">
        <a href="./uid17.html">Highlights of the Year</a>
      </div>
      <div class="TdmEntry">New Software and Platforms<ul><li class="tdmActPage"><a href="uid21.html&#10;&#9;&#9;  ">NetVLAD: CNN architecture for weakly supervised place recognition</a></li><li><a href="uid22.html&#10;&#9;&#9;  ">Unsupervised learning from narrated instruction videos</a></li><li><a href="uid23.html&#10;&#9;&#9;  ">ContextLocNet: Context-aware deep network models for
weakly supervised localization</a></li><li><a href="uid24.html&#10;&#9;&#9;  ">Long-term Temporal Convolutions for Action Recognition</a></li></ul></div>
      <div class="TdmEntry">New Results<ul><li><a href="uid26.html&#10;&#9;&#9;  ">3D object and scene modeling, analysis, and retrieval</a></li><li><a href="uid37.html&#10;&#9;&#9;  ">Category-level object and scene recognition</a></li><li><a href="uid45.html&#10;&#9;&#9;  ">Image restoration, manipulation and enhancement</a></li><li><a href="uid48.html&#10;&#9;&#9;  ">Human activity capture and classification</a></li></ul></div>
      <div class="TdmEntry">Bilateral Contracts and Grants with Industry<ul><li><a href="uid59.html&#10;&#9;&#9;  ">Facebook AI Research Paris: Weakly-supervised interpretation of image and video data (Inria)</a></li><li><a href="uid60.html&#10;&#9;&#9;  ">Google: Learning to annotate videos from movie scripts (Inria)</a></li><li><a href="uid61.html&#10;&#9;&#9;  ">Google: Structured learning from video and natural language (Inria)</a></li><li><a href="uid62.html&#10;&#9;&#9;  ">MSR-Inria joint lab: Image and video mining for science and humanities (Inria)</a></li></ul></div>
      <div class="TdmEntry">Partnerships and Cooperations<ul><li><a href="uid64.html&#10;&#9;&#9;  ">National Initiatives</a></li><li><a href="uid66.html&#10;&#9;&#9;  ">European Initiatives</a></li><li><a href="uid70.html&#10;&#9;&#9;  ">International Initiatives</a></li><li><a href="uid73.html&#10;&#9;&#9;  ">International Research Visitors</a></li></ul></div>
      <div class="TdmEntry">Dissemination<ul><li><a href="uid77.html&#10;&#9;&#9;  ">Promoting Scientific Activities</a></li><li><a href="uid144.html&#10;&#9;&#9;  ">Teaching - Supervision - Juries</a></li><li><a href="uid177.html&#10;&#9;&#9;  ">Popularization</a></li></ul></div>
      <div class="TdmEntry">
        <div>Bibliography</div>
      </div>
      <div class="TdmEntry">
        <ul>
          <li>
            <a id="tdmbibentyear" href="bibliography.html">Publications of the year</a>
          </li>
        </ul>
      </div>
    </div>
    <div id="main">
      <div class="mainentete">
        <div id="head_agauche">
          <small><a href="http://www.inria.fr">
	    
	    Inria
	  </a> | <a href="../index.html">
	    
	    Raweb 
	    2016</a> | <a href="http://www.inria.fr/en/teams/willow">Presentation of the Project-Team WILLOW</a> | <a href="http://www.di.ens.fr/willow">WILLOW Web Site
	  </a></small>
        </div>
        <div id="head_adroite">
          <table class="qrcode">
            <tr>
              <td>
                <a href="willow.xml">
                  <img style="align:bottom; border:none" alt="XML" src="../static/img/icons/xml_motif.png"/>
                </a>
              </td>
              <td>
                <a href="willow.pdf">
                  <img style="align:bottom; border:none" alt="PDF" src="IMG/qrcode-willow-pdf.png"/>
                </a>
              </td>
              <td>
                <a href="../willow/willow.epub">
                  <img style="align:bottom; border:none" alt="e-pub" src="IMG/qrcode-willow-epub.png"/>
                </a>
              </td>
            </tr>
            <tr>
              <td/>
              <td>PDF
</td>
              <td>e-Pub
</td>
            </tr>
          </table>
        </div>
      </div>
      <!--FIN du corps du module-->
      <br/>
      <div class="bottomNavigation">
        <div class="tail_aucentre">
          <a href="./uid17.html" accesskey="P"><img style="align:bottom; border:none" alt="previous" src="../static/img/icons/previous_motif.jpg"/> Previous | </a>
          <a href="./uid0.html" accesskey="U"><img style="align:bottom; border:none" alt="up" src="../static/img/icons/up_motif.jpg"/>  Home</a>
          <a href="./uid22.html" accesskey="N"> | Next <img style="align:bottom; border:none" alt="next" src="../static/img/icons/next_motif.jpg"/></a>
        </div>
        <br/>
      </div>
      <div id="textepage">
        <!--DEBUT2 du corps du module-->
        <h2>Section: 
      New Software and Platforms</h2>
        <h3 class="titre3">NetVLAD: CNN architecture for weakly supervised place recognition</h3>
        <p>Open source release of the software package for our paper "NetVLAD: CNN architecture for weakly supervised place recognition" <a href="./bibliography.html#willow-2016-bid0">[9]</a>. It provides a full implementation of the method, including code for weakly supervised training of the CNN representation, testing on standard datasets, as well as trained models. Links to all of these are available at our project page <a href="http://www.di.ens.fr/willow/research/netvlad/">http://www.di.ens.fr/willow/research/netvlad/</a>.</p>
      </div>
      <!--FIN du corps du module-->
      <br/>
      <div class="bottomNavigation">
        <div class="tail_aucentre">
          <a href="./uid17.html" accesskey="P"><img style="align:bottom; border:none" alt="previous" src="../static/img/icons/previous_motif.jpg"/> Previous | </a>
          <a href="./uid0.html" accesskey="U"><img style="align:bottom; border:none" alt="up" src="../static/img/icons/up_motif.jpg"/>  Home</a>
          <a href="./uid22.html" accesskey="N"> | Next <img style="align:bottom; border:none" alt="next" src="../static/img/icons/next_motif.jpg"/></a>
        </div>
        <br/>
      </div>
    </div>
  </body>
</html>
