<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=utf-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <p><span
        style="font-size:16.0pt;mso-bidi-font-size:12.0pt;mso-bidi-font-weight:
        bold" lang="EN-US">Ph.D position in</span><i
        style="mso-bidi-font-style:normal"><span
          style="font-size:16.0pt; mso-bidi-font-size:12.0pt"
          lang="EN-US"> Material classification based on visual
          appearance</span></i><span
        style="font-size:16.0pt;mso-bidi-font-size:12.0pt;mso-bidi-font-weight:
        bold" lang="EN-US"></span> </p>
    <p
style="margin-top:6.0pt;margin-right:0cm;margin-bottom:0cm;margin-left:0cm;margin-bottom:.0001pt;text-align:justify"><span
        style="mso-bidi-font-weight: bold" lang="EN-US">The Image
        Science and Computer Vision team of Hubert Curien laboratory
        (<a class="moz-txt-link-freetext"
          href="https://laboratoirehubertcurien.univ-st-etienne.fr/en/index.html">https://laboratoirehubertcurien.univ-st-etienne.fr/en/index.html</a>)
        is looking for candidates for a Ph.D position on </span><i
        style="mso-bidi-font-style: normal"><span lang="EN-US">Transfer
          Learning for Material classification based on visual
          appearance correspondences</span></i><span
        style="mso-bidi-font-weight: bold" lang="EN-US">.</span></p>
    <p
style="margin-top:6.0pt;margin-right:0cm;margin-bottom:0cm;margin-left:0cm;margin-bottom:.0001pt;text-align:justify"><span
        style="mso-bidi-font-weight: bold" lang="EN-US">Image
        classification has received a lot of interest in the last decade
        and huge improvements have been observed in terms of
        classification accuracy for the classical datasets such as
        PASCAL VOC or ImageNet. Nevertheless, it appears that material
        classification is still an open problem because of the high
        variability of their appearance in images and because of the
        lack of learning data. In order to cope with these problems,
        recent papers resort to convolutional networks (</span><span
        lang="EN-US"><a
          href="https://arxiv.org/ftp/arxiv/papers/1710/1710.06854.pdf"><span
style="font-size:10.0pt;font-family:"Arial",sans-serif">https://arxiv.org/ftp/arxiv/papers/1710/1710.06854.pdf</span></a><span
          style="mso-bidi-font-weight:bold">) in order to learn the
          variability as well as transfer learning approaches in order
          to be able to learn on different datasets and so increasing
          the amount of learning data (</span><a
          href="https://arxiv.org/pdf/1609.06188.pdf"><span
            style="font-size:10.0pt;
            font-family:"Arial",sans-serif">https://arxiv.org/pdf/1609.06188.pdf</span></a><span
          style="mso-bidi-font-weight:bold">)</span></span><span
        style="font-size:10.0pt;font-family:"Arial",sans-serif"
        lang="EN-US">.</span><span style="mso-bidi-font-weight:bold"
        lang="EN-US"></span></p>
    <p
style="margin-top:6.0pt;margin-right:0cm;margin-bottom:0cm;margin-left:0cm;margin-bottom:.0001pt;text-align:justify"><span
        style="mso-bidi-font-weight: bold" lang="EN-US">The aim of this
        PhD project is to study the visual appearance of materials from
        a computer vision perspective by combining computer vision
        techniques with machine learning and data mining techniques.
        More and more the design of new materials having specific visual
        appearance properties passes through the use of computer based
        approaches (see ref 3, 4 and 5).</span></p>
    <p
style="margin-top:6.0pt;margin-right:0cm;margin-bottom:0cm;margin-left:0cm;margin-bottom:.0001pt;text-align:justify"><span
        style="mso-bidi-font-weight: bold" lang="EN-US">The objective
        will be:</span></p>
    <p
style="margin-top:6.0pt;margin-right:0cm;margin-bottom:5.0pt;margin-left:35.7pt;text-align:justify;text-indent:-17.85pt;mso-list:l1
      level1 lfo2"><span style="mso-bidi-font-weight:bold" lang="EN-US"><span
          style="mso-list:Ignore">1.<span style="font:7.0pt "Times
            New Roman"">      </span></span></span><span dir="LTR"></span><span
        style="mso-bidi-font-weight:bold" lang="EN-US">To study
        different strategies to fuse/combine different datasets, to
        enrich existing datasets using data augmentation methods (e.g.
        light variations, scale, shadows, …), to transfer knowledge
        learnt from one dataset to another one (e.g. see </span><span
        lang="EN-US"><a href="https://arxiv.org/pdf/1609.06188.pdf">https://arxiv.org/pdf/1609.06188.pdf</a>)<span
          style="mso-bidi-font-weight:bold">, to mind/infer knowledge
          from data, etc.</span></span></p>
    <p
style="margin-left:36.0pt;text-align:justify;text-indent:-18.0pt;mso-list:l1
      level1 lfo2"><span style="mso-bidi-font-weight: bold" lang="EN-US"><span
          style="mso-list:Ignore">2.<span style="font:7.0pt "Times
            New Roman"">      </span></span></span><span dir="LTR"></span><span
        style="mso-bidi-font-weight:bold" lang="EN-US">To create a new
        dataset of images of materials which could be complementary to
        the existing synthetized and real-world ones: </span><span
        lang="EN-US">Flickr Material Database (Sharan et al., 2010), the
        ImageNet7 dataset (Hu et al., 2011), the MINC-2500 (Bell et al.,
        2015), the University of Bonn synthetic dataset (Weinmann et
        al., 2014), …<span style="mso-bidi-font-weight:bold"></span></span></p>
    <p
style="margin-left:36.0pt;text-align:justify;text-indent:-18.0pt;mso-list:l1
      level1 lfo2"><span style="mso-bidi-font-weight: bold" lang="EN-US"><span
          style="mso-list:Ignore">3.<span style="font:7.0pt "Times
            New Roman"">      </span></span></span><span dir="LTR"></span><span
        style="mso-bidi-font-weight:bold" lang="EN-US">To classify
        images of materials according their visual appearance in order
        to infer/learn new knowledge on material properties (for example
        using auto-encoders, see <a
          href="https://arxiv.org/pdf/1711.03678.pdf">https://arxiv.org/pdf/1711.03678.pdf</a>).
        Several machine learning and data mining methods (e.g. CNN, deep
        learning, will be investigated. </span></p>
    <p
style="margin-top:5.0pt;margin-right:0cm;margin-bottom:6.0pt;margin-left:35.7pt;text-align:justify;text-indent:-17.85pt;mso-list:l1
      level1 lfo2"><span style="mso-bidi-font-weight:bold" lang="EN-US"><span
          style="mso-list:Ignore">4.<span style="font:7.0pt "Times
            New Roman"">      </span></span></span><span dir="LTR"></span><span
        style="mso-bidi-font-weight:bold" lang="EN-US">To learn how to
        characterize the visual appearance of some materials from a
        limited set of features and of image acquisitions. The
        auto-encoder could be a nice tool to access semantic features
        and observe their impact on the reconstructed material images.
        This could also help for material design.</span></p>
    <p style="margin-top:6.0pt;text-align:justify"><span
        style="mso-bidi-font-weight:bold" lang="EN-US">The thesis will
        be co-supervised by Alain Trémeau (Full Professor, <a
          href="https://perso.univ-st-etienne.fr/tremeaua/">https://perso.univ-st-etienne.fr/tremeaua/</a>)
        and Damien Muselet (Assistant Professor,</span><span
        lang="EN-US"> <span style="mso-bidi-font-weight:bold"><a
            href="https://perso.univ-st-etienne.fr/muda8804/">https://perso.univ-st-etienne.fr/muda8804/</a>). </span></span></p>
    <p style="text-align:justify"><b style="mso-bidi-font-weight:normal"><span
          lang="EN-US">The deadline for applications is 06/05/2018.</span></b></p>
    <p class="MsoNormal"><span style="font-size:6.0pt;line-height:107%;
        font-family:"Times New Roman",serif" lang="EN-US"> </span></p>
    <p class="MsoNormal"><b style="mso-bidi-font-weight:normal"><span
          style="font-size:16.0pt;line-height:107%;font-family:"Times
          New Roman",serif; mso-ansi-language:FR">Bibliography</span></b></p>
    <p class="MsoListParagraphCxSpFirst"
      style="margin-top:6.0pt;margin-right:0cm;
margin-bottom:6.0pt;margin-left:35.7pt;mso-add-space:auto;text-indent:-17.85pt;
      mso-list:l0 level1 lfo1"><span class="MsoHyperlink"><span
          style="color:#0563C1;mso-ansi-language:EN-US;text-decoration:none;
          text-underline:none" lang="EN-US"><span
            style="mso-list:Ignore">1.<span style="font:7.0pt
              "Times New Roman"">      </span></span></span></span><span
        dir="LTR"></span><span style="mso-ansi-language:EN-US"
        lang="EN-US">Sébastien Lagarde, “Open Problems in Real-Time
        Rendering-Physically-Based Materials: Where Are We?” in ACM
        SIGGRAPH 2017, </span><a
href="HTTP://openproblems.realtimerendering.com/s2017/02-PhysicallyBasedMaterialWhereAreWe.pdf"><span
          style="mso-ansi-language:EN-US" lang="EN-US">http://openproblems.realtimerendering.com/s2017/02-PhysicallyBasedMaterialWhereAreWe.pdf</span></a><span
        class="MsoHyperlink"><span
          style="color:#0563C1;mso-ansi-language: EN-US" lang="EN-US"></span></span></p>
    <p class="MsoListParagraphCxSpMiddle"
      style="margin-top:6.0pt;margin-right:0cm;
margin-bottom:6.0pt;margin-left:35.7pt;mso-add-space:auto;text-indent:-17.85pt;
      mso-list:l0 level1 lfo1"><span
        style="color:#0563C1;mso-ansi-language:EN-US" lang="EN-US"><span
          style="mso-list:Ignore">2.<span style="font:7.0pt "Times
            New Roman"">      </span></span></span><span dir="LTR"></span><span
        style="letter-spacing:.1pt;mso-ansi-language: EN-US"
        lang="EN-US">(2018) </span><span
        style="mso-ansi-language:EN-US" lang="EN-US">G. Kalliatakis, A.
        Sticlaru, G. Stamatiadis, S. Ehsan, A. Leonardis, J. Gall and K.
        D. McDonald-Maier, Material Classification in the Wild: Do
        Synthesized Training Data Generalise Better than Real-world
        Training Data?<span style="mso-spacerun:yes">  </span>Proceedings
        of VISAPP’2018.<u><span style="color:#0563C1"></span></u></span></p>
    <p class="MsoListParagraphCxSpMiddle"
      style="margin-top:6.0pt;margin-right:0cm;
margin-bottom:6.0pt;margin-left:35.7pt;mso-add-space:auto;text-indent:-17.85pt;
      mso-list:l0 level1 lfo1"><span class="LienInternet"><span
          style="mso-ansi-language:EN-US;text-decoration:none;text-underline:
          none" lang="EN-US"><span style="mso-list:Ignore">3.<span
              style="font:7.0pt "Times New Roman"">      </span></span></span></span><span
        dir="LTR"></span><span
        style="letter-spacing:.1pt;mso-ansi-language:EN-US" lang="EN-US">(2018)
        Reviewing the Novel Machine Learning Tools for Materials Design.
      </span><a
        href="https://link.springer.com/chapter/10.1007/978-3-319-67459-9_7"><span
          style="mso-ansi-language:EN-US" lang="EN-US">https://link.springer.com/chapter/10.1007/978-3-319-67459-9_7</span></a><span
        class="LienInternet"><span style="mso-ansi-language:EN-US"
          lang="EN-US">; </span></span></p>
    <p class="MsoListParagraphCxSpMiddle"
      style="margin-top:6.0pt;margin-right:0cm;
margin-bottom:6.0pt;margin-left:35.7pt;mso-add-space:auto;text-indent:-17.85pt;
      mso-list:l0 level1 lfo1"><span class="LienInternet"><span
          style="mso-ansi-language:EN-US;text-decoration:none;text-underline:
          none" lang="EN-US"><span style="mso-list:Ignore">4.<span
              style="font:7.0pt "Times New Roman"">      </span></span></span></span><span
        dir="LTR"></span><span
        style="letter-spacing:.1pt;mso-ansi-language:EN-US" lang="EN-US">(2017)
      </span><span style="mso-ansi-language:EN-US" lang="EN-US">Data
        mining-aided materials discovery and optimization,<span
          style="mso-spacerun:yes">  </span></span><a
href="http://www.sciencedirect.com/science/article/pii/S2352847817300618"><span
          class="LienInternet"><span
            style="color:blue;mso-ansi-language:EN-US" lang="EN-US">http://www.sciencedirect.com/science/article/pii/S2352847817300618</span></span></a><span
        class="LienInternet"><span style="mso-ansi-language:EN-US"
          lang="EN-US">; </span></span></p>
    <p class="MsoListParagraphCxSpMiddle"
      style="margin-top:6.0pt;margin-right:0cm;
margin-bottom:6.0pt;margin-left:35.7pt;mso-add-space:auto;text-indent:-17.85pt;
      mso-list:l0 level1 lfo1"><span class="LienInternet"><span
          style="mso-ansi-language:EN-US;text-decoration:none;text-underline:
          none" lang="EN-US"><span style="mso-list:Ignore">5.<span
              style="font:7.0pt "Times New Roman"">      </span></span></span></span><span
        dir="LTR"></span><span class="text"><span
          style="mso-ansi-language:EN-US" lang="EN-US">(2017) </span></span><span
        class="author-ref"><sup><span style="mso-ansi-language:EN-US"
            lang="EN-US"><span style="mso-spacerun:yes"> </span></span></sup></span><span
        style="mso-ansi-language:EN-US" lang="EN-US">Materials discovery
        and design using machine learning, </span><a
href="http://www.sciencedirect.com/science/article/pii/S2352847817300515"><span
          style="mso-ansi-language:EN-US" lang="EN-US">http://www.sciencedirect.com/science/article/pii/S2352847817300515</span></a><span
        class="LienInternet"><span style="mso-ansi-language:EN-US"
          lang="EN-US">; </span></span></p>
    <p class="MsoListParagraphCxSpLast"
      style="margin-top:6.0pt;margin-right:0cm;
margin-bottom:6.0pt;margin-left:35.7pt;mso-add-space:auto;text-indent:-17.85pt;
      mso-list:l0 level1 lfo1"><span
        style="color:#0563C1;mso-ansi-language:EN-US" lang="EN-US"><span
          style="mso-list:Ignore">6.<span style="font:7.0pt "Times
            New Roman"">      </span></span></span><span dir="LTR"></span><span
        style="mso-ansi-language:EN-US;mso-fareast-language: ZH-TW"
        lang="EN-US">(2016) An intuitive control space for material
        appearanc<span style="mso-bidi-font-weight:bold">e<b>, </b></span></span><a
        href="https://dl.acm.org/citation.cfm?id=2980242"><span
          style="mso-ansi-language:EN-US" lang="EN-US">https://dl.acm.org/citation.cfm?id=2980242</span></a><u><span
          style="color:#0563C1;mso-ansi-language:EN-US" lang="EN-US"></span></u></p>
    <h2 style="margin-bottom:6.0pt"><span style="font-size:16.0pt"
        lang="EN-US">Requested skills</span></h2>
    <p style="margin-top:6.0pt;text-align:justify"><span
        style="mso-bidi-font-weight:bold" lang="EN-US">The desired
        profile is Master (MSc or equivalent) or Engineer degree in
        Machine Learning and Data Mining / Image Processing and Computer
        Vision / Computer Science and Applied Mathematics, with
        excellent academic record and research experience, in-depth
        knowledge of machine learning (Computational Neural Networks,
        Deep Learning), data mining (Transfer Knowledge), optimization
        methods, with a specialization in one of the following areas:
        machine learning, data mining or computer vision.</span></p>
    <p style="margin-top:6.0pt;text-align:justify"><span lang="EN-US">We
        are looking for a curious student with excellent programming
        skills (e.g., in Matlab, Python, or C/C++).<span
          style="mso-bidi-font-weight:bold"></span></span></p>
    <p class="MsoNormal"
      style="mso-margin-top-alt:auto;margin-bottom:6.0pt;
      line-height:normal;mso-outline-level:2"><b><span style="font-size:
          16.0pt;font-family:"Times New
          Roman",serif;mso-fareast-font-family:"Times New
          Roman"" lang="EN-US">Application</span></b></p>
    <p style="margin-top:6.0pt"><span lang="EN-US">Interested candidates
        should send a resume, a cover letter, and transcripts of BSc and
        MSc (M1 and M2 years). Recommendation letters will be
        appreciated.</span></p>
    <p style="margin-top:6.0pt"><span lang="EN-US">All applications must
        be sent electronically to Alain Trémeau (<a
          href="mailto:alain.tremeau@univ-st-etienne.fr">alain.tremeau@univ-st-etienne.fr</a>)
        and Damien Muselet (<a
          href="mailto:damien.muselet@univ-st-etienne.fr">damien.muselet@univ-st-etienne.fr</a>)
      </span></p>
    <h2 style="margin-bottom:6.0pt"><span style="font-size:16.0pt"
        lang="EN-US">Contract</span></h2>
    <p style="margin-top:6.0pt"><span lang="EN-US">3-years contract on
        the basis of a monthly gross income of 1 760 euros
        approximatively. Part-time teaching can be considered. Start in
        autumn 2018.</span></p>
    <div class="moz-signature">-- <br>
      <meta charset="UTF-8">
      <br>
      <br>
      <table style="background-color: #FFFFFF; border-collapse:
        collapse;">
        <tbody>
          <tr style="height: 120px">
            <td style="vertical-align: top; padding:0px; padding-right:
              10px; border-right: 1px solid #E9540D"><a
                href="http://laboratoirehubertcurien.fr"><img
                  src="cid:part15.90DC1639.F218C07F@univ-st-etienne.fr"></a></td>
            <td style="vertical-align: top; padding:0px; padding-top:
              3px; padding-left: 10px; padding-right: 10px;
              border-right: 1px solid #E9540D;">
              <table style="font: 10pt 'Tahoma', sans-serif;">
                <tbody>
                  <tr>
                    <td style="font-weight: 700; color:#000000;
                      font-size: 10pt;">Alain <span
                        style="text-transform:uppercase;">Tremeau</span></td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">Professor</td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">Academic Coordinator of
                      Masters COSI/CIMET and 3DMT,
                      <a class="moz-txt-link-freetext" href="http://www.master-colorscience.eu/">http://www.master-colorscience.eu/</a>,
                      <a class="moz-txt-link-freetext" href="http://master-3dmt.eu/">http://master-3dmt.eu/</a></td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">Visit my homepage:
                      <a class="moz-txt-link-freetext" href="http://perso.univ-st-etienne.fr/tremeaua/">http://perso.univ-st-etienne.fr/tremeaua/</a></td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D"><a style="text-decoration:
                        inherit; color: inherit;"
                        href="mailto:alain.tremeau@univ-st-etienne.fr">alain.tremeau@univ-st-etienne.fr</a></td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">Tél. : 04 77 91 57 52</td>
                  </tr>
                </tbody>
              </table>
            </td>
            <td style="vertical-align: top; padding:0px; padding-top:
              3px; padding-left: 10px; padding-right: 10px;
              border-right: 1px solid #E9540D;">
              <table style="font: 10pt 'Tahoma', sans-serif;">
                <tbody>
                  <tr>
                    <td style="font-weight: 700; color:#000000;
                      font-size: 10pt;">Laboratoire Hubert Curien</td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">Image Science &
                      Computer Vision Group</td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">Campus Manufacture</td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">23 RUE Dr Paul Michelon</td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">42023 SAINT-ETIENNE CEDEX
                      2</td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D">04 77 91 57 80</td>
                  </tr>
                  <tr>
                    <td style="color:#E9540D"><a style="text-decoration:
                        inherit; color: inherit;"
                        href="http://laboratoirehubertcurien.fr">http://laboratoirehubertcurien.fr</a></td>
                  </tr>
                </tbody>
              </table>
            </td>
            <td style="vertical-align: top; padding:0px; padding-left:
              10px; padding-right: 10px; border-right: 1px solid
              #E9540D;">
              <table>
                <tbody>
                  <tr>
                    <td colspan="5"><a
                        href="http://www.univ-st-etienne.fr"><img
                          src="cid:part19.0E5941FA.DBD2AB12@univ-st-etienne.fr"></a></td>
                  </tr>
                  <tr>
                  </tr>
                  <tr>
                    <td style="margin-right: 4px;"><a
                        style="border-style: none;"
                        href="http://www.facebook.com/Universite.Jean.Monnet.Saint.Etienne"><img
src="cid:part21.7E76A31C.3B4D4F14@univ-st-etienne.fr"></a></td>
                    <td style="margin-right: 4px;"><a
                        style="border-style: none;"
                        href="https://twitter.com/Univ_St_Etienne"><img
src="cid:part23.B45C7A88.0A99C315@univ-st-etienne.fr"></a></td>
                    <td style="margin-right: 4px;"><a
                        style="border-style: none;"
                        href="https://www.youtube.com/user/UniJeanMonnetUJM"><img
src="cid:part25.431C42E5.D632E5EB@univ-st-etienne.fr"></a></td>
                    <td style="margin-right: 4px;"><a
                        style="border-style: none;"
href="https://www.linkedin.com/edu/universit%C3%A9-jean-monnet-saint-etienne-12533"><img
src="cid:part27.B8D3F196.26BE4E33@univ-st-etienne.fr"></a></td>
                    <td style="margin-right: 4px;"><a
                        style="border-style: none; text-decoration:
                        none;"
                        href="http://www.viadeo.com/fr/company/universite-jean-monnet"><img
src="cid:part29.630A82E8.5953930E@univ-st-etienne.fr"></a></td>
                  </tr>
                </tbody>
              </table>
            </td>
            <td style="vertical-align: top; padding:0px; padding-top:
              3px; padding-left: 10px; padding-right: 10px;"> <a
                href="http://www.universite-lyon.fr"><img
                  src="cid:part31.BB6A7900.F86C1153@univ-st-etienne.fr"></a>
            </td>
          </tr>
        </tbody>
      </table>
    </div>
  </body>
</html>