<?xml version="1.0" encoding="utf-8"?>
<TEI xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:hal="http://hal.archives-ouvertes.fr/" xmlns:gml="http://www.opengis.net/gml/3.3/" xmlns:gmlce="http://www.opengis.net/gml/3.3/ce" version="1.1" xsi:schemaLocation="http://www.tei-c.org/ns/1.0 http://api.archives-ouvertes.fr/documents/aofr-sword.xsd">
  <teiHeader>
    <fileDesc>
      <titleStmt>
        <title>HAL TEI export of hal-03481207</title>
      </titleStmt>
      <publicationStmt>
        <distributor>CCSD</distributor>
        <availability status="restricted">
          <licence target="https://creativecommons.org/publicdomain/zero/1.0/">CC0 1.0 - Universal</licence>
        </availability>
        <date when="2026-05-17T03:07:29+02:00"/>
      </publicationStmt>
      <sourceDesc>
        <p part="N">HAL API Platform</p>
      </sourceDesc>
    </fileDesc>
  </teiHeader>
  <text>
    <body>
      <listBibl>
        <biblFull>
          <titleStmt>
            <title xml:lang="en">Segmenter: Transformer for Semantic Segmentation</title>
            <author role="aut">
              <persName>
                <forename type="first">Robin</forename>
                <surname>Strudel</surname>
              </persName>
              <idno type="halauthorid">1641541-0</idno>
              <affiliation ref="#struct-454692"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Ricardo</forename>
                <surname>Garcia</surname>
              </persName>
              <email type="md5">15a0964065b6f6e60e3cf428d8c1ade4</email>
              <email type="domain">inria.fr</email>
              <idno type="idhal" notation="string">ricardo-garcia-pinel</idno>
              <idno type="idhal" notation="numeric">1653386</idno>
              <idno type="halauthorid" notation="string">39487-1653386</idno>
              <idno type="ORCID">https://orcid.org/0000-0002-2553-7272</idno>
              <idno type="GOOGLE SCHOLAR">cMA5vJwAAAAJ</idno>
              <affiliation ref="#struct-454692"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Ivan</forename>
                <surname>Laptev</surname>
              </persName>
              <idno type="halauthorid">319698-0</idno>
              <affiliation ref="#struct-454692"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Cordelia</forename>
                <surname>Schmid</surname>
              </persName>
              <idno type="halauthorid">119565-0</idno>
              <affiliation ref="#struct-454692"/>
            </author>
            <editor role="depositor">
              <persName>
                <forename>Robin</forename>
                <surname>Strudel</surname>
              </persName>
              <email type="md5">b9349a4f9b1390da2e0f74841a9e237c</email>
              <email type="domain">gmail.com</email>
            </editor>
            <funder ref="#projanr-50388"/>
            <funder>This work was partially supported by the HPC resources from GENCI-IDRIS (Grant 2020-AD011011163R1), the Louis Vuitton ENS Chair on Artificial Intelligence, and the French government under man-agement of Agence Nationale de la Recherche as part of the ”Investissements d’avenir” program, reference ANR-19-P3IA-0001 (PRAIRIE 3IA Institute).</funder>
          </titleStmt>
          <editionStmt>
            <edition n="v1" type="current">
              <date type="whenSubmitted">2021-12-15 11:03:05</date>
              <date type="whenModified">2026-03-04 03:07:20</date>
              <date type="whenReleased">2021-12-15 11:24:29</date>
              <date type="whenProduced">2021-10-11</date>
              <date type="whenEndEmbargoed">2021-12-15</date>
              <ref type="file" target="https://hal.science/hal-03481207v1/document">
                <date notBefore="2021-12-15"/>
              </ref>
              <ref type="file" subtype="author" n="1" target="https://hal.science/hal-03481207v1/file/2105.05633.pdf" id="file-3481207-3053004">
                <date notBefore="2021-12-15"/>
              </ref>
            </edition>
            <respStmt>
              <resp>contributor</resp>
              <name key="823130">
                <persName>
                  <forename>Robin</forename>
                  <surname>Strudel</surname>
                </persName>
                <email type="md5">b9349a4f9b1390da2e0f74841a9e237c</email>
                <email type="domain">gmail.com</email>
              </name>
            </respStmt>
          </editionStmt>
          <publicationStmt>
            <distributor>CCSD</distributor>
            <idno type="halId">hal-03481207</idno>
            <idno type="halUri">https://hal.science/hal-03481207</idno>
            <idno type="halBibtex">strudel:hal-03481207</idno>
            <idno type="halRefHtml">&lt;i&gt;ICCV 2021 - International Conference on Computer Vision&lt;/i&gt;, Oct 2021, Virtual, France</idno>
            <idno type="halRef">ICCV 2021 - International Conference on Computer Vision, Oct 2021, Virtual, France</idno>
            <availability status="restricted">
              <licence target="https://about.hal.science/hal-authorisation-v1/">HAL Authorization<ref corresp="#file-3481207-3053004"/></licence>
            </availability>
          </publicationStmt>
          <seriesStmt>
            <idno type="stamp" n="ENS-PARIS">Ecole Normale Supérieure de Paris</idno>
            <idno type="stamp" n="CNRS">CNRS - Centre national de la recherche scientifique</idno>
            <idno type="stamp" n="INRIA">INRIA - Institut National de Recherche en Informatique et en Automatique</idno>
            <idno type="stamp" n="INRIA-ROCQ">INRIA Paris - Rocquencourt</idno>
            <idno type="stamp" n="TESTALAIN1">TESTALAIN1</idno>
            <idno type="stamp" n="INRIA2">INRIA 2</idno>
            <idno type="stamp" n="PSL">Université Paris sciences et lettres</idno>
            <idno type="stamp" n="INRIA-PSL">INRIA-PSL</idno>
            <idno type="stamp" n="TEST-HALCNRS">Collection test HAL CNRS</idno>
            <idno type="stamp" n="PNRIA">Programme National de Recherche en IA</idno>
            <idno type="stamp" n="ENS-PSL" corresp="PSL">École normale supérieure - PSL</idno>
            <idno type="stamp" n="ANR">ANR</idno>
            <idno type="stamp" n="PRAIRIE-IA">PaRis AI Research InstitutE</idno>
            <idno type="stamp" n="DIENS" corresp="ENS-PARIS">Département d'informatique de l'ENS-PSL</idno>
            <idno type="stamp" n="ANR-IA-19" corresp="ANR-IA">ANR-IA-19</idno>
            <idno type="stamp" n="ANR-IA">ANR-IA</idno>
          </seriesStmt>
          <notesStmt>
            <note type="audience" n="2">International</note>
            <note type="invited" n="0">No</note>
            <note type="popular" n="0">No</note>
            <note type="peer" n="1">Yes</note>
            <note type="proceedings" n="1">Yes</note>
          </notesStmt>
          <sourceDesc>
            <biblStruct>
              <analytic>
                <title xml:lang="en">Segmenter: Transformer for Semantic Segmentation</title>
                <author role="aut">
                  <persName>
                    <forename type="first">Robin</forename>
                    <surname>Strudel</surname>
                  </persName>
                  <idno type="halauthorid">1641541-0</idno>
                  <affiliation ref="#struct-454692"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Ricardo</forename>
                    <surname>Garcia</surname>
                  </persName>
                  <email type="md5">15a0964065b6f6e60e3cf428d8c1ade4</email>
                  <email type="domain">inria.fr</email>
                  <idno type="idhal" notation="string">ricardo-garcia-pinel</idno>
                  <idno type="idhal" notation="numeric">1653386</idno>
                  <idno type="halauthorid" notation="string">39487-1653386</idno>
                  <idno type="ORCID">https://orcid.org/0000-0002-2553-7272</idno>
                  <idno type="GOOGLE SCHOLAR">cMA5vJwAAAAJ</idno>
                  <affiliation ref="#struct-454692"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Ivan</forename>
                    <surname>Laptev</surname>
                  </persName>
                  <idno type="halauthorid">319698-0</idno>
                  <affiliation ref="#struct-454692"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Cordelia</forename>
                    <surname>Schmid</surname>
                  </persName>
                  <idno type="halauthorid">119565-0</idno>
                  <affiliation ref="#struct-454692"/>
                </author>
              </analytic>
              <monogr>
                <meeting>
                  <title>ICCV 2021 - International Conference on Computer Vision</title>
                  <date type="start">2021-10-11</date>
                  <date type="end">2021-10-17</date>
                  <settlement>Virtual</settlement>
                  <country key="FR">France</country>
                </meeting>
                <imprint/>
              </monogr>
            </biblStruct>
          </sourceDesc>
          <profileDesc>
            <langUsage>
              <language ident="en">English</language>
            </langUsage>
            <textClass>
              <classCode scheme="halDomain" n="info.info-ai">Computer Science [cs]/Artificial Intelligence [cs.AI]</classCode>
              <classCode scheme="halDomain" n="info">Computer Science [cs]</classCode>
              <classCode scheme="halDomain" n="info.info-cv">Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV]</classCode>
              <classCode scheme="halDomain" n="info.info-lg">Computer Science [cs]/Machine Learning [cs.LG]</classCode>
              <classCode scheme="halTypology" n="COMM">Conference papers</classCode>
              <classCode scheme="halOldTypology" n="COMM">Conference papers</classCode>
              <classCode scheme="halTreeTypology" n="COMM">Conference papers</classCode>
            </textClass>
            <abstract xml:lang="en">
              <p>Image segmentation is often ambiguous at the level of individual image patches and requires contextual information to reach label consensus. In this paper we introduce Segmenter, a transformer model for semantic segmentation. In contrast to convolution-based methods, our approach allows to model global context already at the first layer and throughout the network. We build on the recent Vision Transformer (ViT) and extend it to semantic segmentation. To do so, we rely on the output embeddings corresponding to image patches and obtain class labels from these embeddings with a point-wise linear decoder or a mask transformer decoder. We leverage models pre-trained for image classification and show that we can fine-tune them on moderate sized datasets available for semantic segmentation. The linear decoder allows to obtain excellent results already, but the performance can be further improved by a mask transformer generating class masks. We conduct an extensive ablation study to show the impact of the different parameters, in particular the performance is better for large models and small patch sizes. Segmenter attains excellent results for semantic segmentation. It outperforms the state of the art on both ADE20K and Pascal Context datasets and is competitive on Cityscapes.</p>
            </abstract>
          </profileDesc>
        </biblFull>
      </listBibl>
    </body>
    <back>
      <listOrg type="structures">
        <org type="researchteam" xml:id="struct-454692" status="VALID">
          <idno type="IdRef">229858406</idno>
          <idno type="RNSR">200718311C</idno>
          <idno type="ROR">https://ror.org/04c96ng72</idno>
          <orgName>Models of visual object recognition and scene understanding</orgName>
          <orgName type="acronym">WILLOW</orgName>
          <date type="start">2016-01-01</date>
          <date type="end">2028-06-30</date>
          <desc>
            <address>
              <addrLine>48 Rue Barrault, 75013 Paris</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.inria.fr/equipes/willow</ref>
          </desc>
          <listRelation>
            <relation active="#struct-25027" type="direct"/>
            <relation active="#struct-59704" type="indirect"/>
            <relation active="#struct-564132" type="indirect"/>
            <relation active="#struct-300009" type="indirect"/>
            <relation name="UMR8548" active="#struct-441569" type="indirect"/>
            <relation active="#struct-454310" type="direct"/>
          </listRelation>
        </org>
        <org type="regrouplaboratory" xml:id="struct-25027" status="VALID">
          <idno type="IdRef">148034055</idno>
          <idno type="RNSR">199812876J</idno>
          <idno type="ROR">05y6rqs46</idno>
          <orgName>Département d'informatique - ENS-PSL</orgName>
          <orgName type="acronym">DI-ENS</orgName>
          <date type="start">1999-01-01</date>
          <desc>
            <address>
              <addrLine>École normale supérieure 45 rue d'Ulm F-75230 Paris Cedex 05</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.di.ens.fr/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-59704" type="direct"/>
            <relation active="#struct-564132" type="indirect"/>
            <relation active="#struct-300009" type="direct"/>
            <relation name="UMR8548" active="#struct-441569" type="direct"/>
          </listRelation>
        </org>
        <org type="institution" xml:id="struct-59704" status="VALID">
          <idno type="IdRef">031738419</idno>
          <idno type="ISNI">0000000123532622</idno>
          <idno type="ROR">https://ror.org/05a0dhs15</idno>
          <orgName>École normale supérieure - Paris</orgName>
          <orgName type="acronym">ENS-PSL</orgName>
          <date type="start">1985-07-24</date>
          <desc>
            <address>
              <addrLine>45, Rue d'Ulm - 75230 Paris cedex 05</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">https://www.ens.psl.eu/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-564132" type="direct"/>
          </listRelation>
        </org>
        <org type="regroupinstitution" xml:id="struct-564132" status="VALID">
          <idno type="IdRef">241597595</idno>
          <idno type="ISNI">0000 0004 1784 3645</idno>
          <idno type="ROR">https://ror.org/013cjyk83</idno>
          <orgName>Université Paris Sciences et Lettres</orgName>
          <orgName type="acronym">PSL</orgName>
          <desc>
            <address>
              <addrLine>60 rue Mazarine 75006 Paris</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">https://www.psl.eu/</ref>
          </desc>
        </org>
        <org type="institution" xml:id="struct-300009" status="VALID">
          <idno type="ROR">https://ror.org/02kvxyf05</idno>
          <orgName>Institut National de Recherche en Informatique et en Automatique</orgName>
          <orgName type="acronym">Inria</orgName>
          <desc>
            <address>
              <addrLine>Domaine de VoluceauRocquencourt - BP 10578153 Le Chesnay Cedex</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.inria.fr/en/</ref>
          </desc>
        </org>
        <org type="regroupinstitution" xml:id="struct-441569" status="VALID">
          <idno type="IdRef">02636817X</idno>
          <idno type="ISNI">0000000122597504</idno>
          <idno type="ROR">https://ror.org/02feahw73</idno>
          <orgName>Centre National de la Recherche Scientifique</orgName>
          <orgName type="acronym">CNRS</orgName>
          <date type="start">1939-10-19</date>
          <desc>
            <address>
              <country key="FR"/>
            </address>
            <ref type="url">https://www.cnrs.fr/</ref>
          </desc>
        </org>
        <org type="laboratory" xml:id="struct-454310" status="VALID">
          <idno type="IdRef">241614864</idno>
          <idno type="RNSR">196718247G</idno>
          <idno type="ROR">https://ror.org/05eyd5d35</idno>
          <orgName>Centre Inria de Paris</orgName>
          <date type="start">2016-03-10</date>
          <desc>
            <address>
              <addrLine>48 Rue Barrault, 75013 Paris</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.inria.fr/centre/paris</ref>
          </desc>
          <listRelation>
            <relation active="#struct-300009" type="direct"/>
          </listRelation>
        </org>
      </listOrg>
      <listOrg type="projects">
        <org type="anrProject" xml:id="projanr-50388" status="VALID">
          <idno type="anr">ANR-19-P3IA-0001</idno>
          <orgName>PRAIRIE</orgName>
          <desc>PaRis Artificial Intelligence Research InstitutE</desc>
          <date type="start">2019</date>
        </org>
      </listOrg>
    </back>
  </text>
</TEI>