The SIFT Method
Identifieur interne : 004407 ( Main/Merge ); précédent : 004406; suivant : 004408The SIFT Method
Auteurs : Frédéric Cao [France] ; José-Luis Lisani [Espagne] ; Jean-Michel Morel [France] ; Pablo Musé [Uruguay] ; Frédéric Sur [France]Source :
- Lecture Notes in Mathematics [ 0075-8434 ]
Abstract
In this chapter and in the next one, we describe one of the most popular shape descriptors, Lowe's Scale-Invariant Feature Transform (SIFT) method [114]. In continuation we will perform a structural and practical comparison of the SIFTbased matching method with the Level Line Descriptor method (LLD) developed in this book. The LLD method in fact includes the features of the recent, also popular, MSER method. Comparing SIFT and LLD is not an easy task, since they are of different nature. On the one hand LLD is based on geometrical shape descriptors, rigorously invariant with respect to similarity or affine transformations. Moreover, the method comes with decision rules, either for matching or grouping. On the other hand, SIFT descriptors are local patches which are based on key points and which are just similarity-invariant. The comparison will be based on ad hoc experimental protocols, in the spirit of the SIFT method itself. These protocols check the robustness of local descriptors to all perturbations listed in Sect. 1.2 (Chap. 1). We start with a comprehensive description of the SIFT shape encoding (Sect. 10.1). Then we compare robustness and stability of both shape descriptors (Sect. 10.2). Sect. 10.3 compares the matching performances of both algorithms on pair of images having similar shapes or obtained by photographing the same scene under different viewpoints. The main focus of the book is the computation of matching thresholds. In the SIFT method the thresholds are learned from the pair of images. We shall see that obvious matches can be missed when the query shape appears more than once in the searched image. In the next chapter a fusion of SIFT and of the a contrario techniques both for matching and grouping will be proposed.
Url:
DOI: 10.1007/978-3-540-68481-7_10
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 001B24
- to stream Istex, to step Curation: 001B04
- to stream Istex, to step Checkpoint: 000D24
Links to Exploration step
ISTEX:763F1BE2D6282001100A5C5613EF3840B2713620Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">The SIFT Method</title>
<author wicri:is="90%"><name sortKey="Cao, Frederic" sort="Cao, Frederic" uniqKey="Cao F" first="Frédéric" last="Cao">Frédéric Cao</name>
</author>
<author wicri:is="90%"><name sortKey="Lisani, Jose Luis" sort="Lisani, Jose Luis" uniqKey="Lisani J" first="José-Luis" last="Lisani">José-Luis Lisani</name>
</author>
<author wicri:is="90%"><name sortKey="Morel, Jean Michel" sort="Morel, Jean Michel" uniqKey="Morel J" first="Jean-Michel" last="Morel">Jean-Michel Morel</name>
</author>
<author wicri:is="90%"><name sortKey="Muse, Pablo" sort="Muse, Pablo" uniqKey="Muse P" first="Pablo" last="Musé">Pablo Musé</name>
</author>
<author wicri:is="90%"><name sortKey="Sur, Frederic" sort="Sur, Frederic" uniqKey="Sur F" first="Frédéric" last="Sur">Frédéric Sur</name>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">ISTEX</idno>
<idno type="RBID">ISTEX:763F1BE2D6282001100A5C5613EF3840B2713620</idno>
<date when="2008" year="2008">2008</date>
<idno type="doi">10.1007/978-3-540-68481-7_10</idno>
<idno type="url">https://api.istex.fr/ark:/67375/HCB-M1WF8XZF-2/fulltext.pdf</idno>
<idno type="wicri:Area/Istex/Corpus">001B24</idno>
<idno type="wicri:explorRef" wicri:stream="Istex" wicri:step="Corpus" wicri:corpus="ISTEX">001B24</idno>
<idno type="wicri:Area/Istex/Curation">001B04</idno>
<idno type="wicri:Area/Istex/Checkpoint">000D24</idno>
<idno type="wicri:explorRef" wicri:stream="Istex" wicri:step="Checkpoint">000D24</idno>
<idno type="wicri:doubleKey">0075-8434:2008:Cao F:the:sift:method</idno>
<idno type="wicri:Area/Main/Merge">004407</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">The SIFT Method</title>
<author wicri:is="90%"><name sortKey="Cao, Frederic" sort="Cao, Frederic" uniqKey="Cao F" first="Frédéric" last="Cao">Frédéric Cao</name>
<affiliation wicri:level="3"><country xml:lang="fr">France</country>
<wicri:regionArea>DxO Labs, 3 rue Nationale, 92100 Boulogne Billancourt</wicri:regionArea>
<placeName><region type="region" nuts="2">Île-de-France</region>
<settlement type="city">Boulogne-Billancourt</settlement>
</placeName>
</affiliation>
<affiliation></affiliation>
</author>
<author wicri:is="90%"><name sortKey="Lisani, Jose Luis" sort="Lisani, Jose Luis" uniqKey="Lisani J" first="José-Luis" last="Lisani">José-Luis Lisani</name>
<affiliation wicri:level="1"><country xml:lang="fr">Espagne</country>
<wicri:regionArea>Dep. Matemàtiques i Informàtica, University Balearic Islands, ctra. Valldemossa km.7,5, 07122 Palma de Mallorca, Balears</wicri:regionArea>
<wicri:noRegion>Balears</wicri:noRegion>
</affiliation>
<affiliation wicri:level="1"><country wicri:rule="url">Espagne</country>
</affiliation>
</author>
<author wicri:is="90%"><name sortKey="Morel, Jean Michel" sort="Morel, Jean Michel" uniqKey="Morel J" first="Jean-Michel" last="Morel">Jean-Michel Morel</name>
<affiliation wicri:level="3"><country xml:lang="fr">France</country>
<wicri:regionArea>Ecole Normale Supérieure de Cachan, CMLA, 61 av. du Président Wilson, 94235 Cachan Cédex</wicri:regionArea>
<placeName><region type="region" nuts="2">Île-de-France</region>
<settlement type="city">Cachan Cédex</settlement>
</placeName>
</affiliation>
<affiliation wicri:level="1"><country wicri:rule="url">France</country>
</affiliation>
</author>
<author wicri:is="90%"><name sortKey="Muse, Pablo" sort="Muse, Pablo" uniqKey="Muse P" first="Pablo" last="Musé">Pablo Musé</name>
<affiliation wicri:level="1"><country xml:lang="fr">Uruguay</country>
<wicri:regionArea>Instituto de Ingeniería Eléctrica, Julio Herrera y Reissig 565, 11300 Montevideo</wicri:regionArea>
<wicri:noRegion>11300 Montevideo</wicri:noRegion>
</affiliation>
<affiliation wicri:level="1"><country wicri:rule="url">Uruguay</country>
</affiliation>
</author>
<author wicri:is="90%"><name sortKey="Sur, Frederic" sort="Sur, Frederic" uniqKey="Sur F" first="Frédéric" last="Sur">Frédéric Sur</name>
<affiliation wicri:level="1"><country xml:lang="fr">France</country>
<wicri:regionArea>Loria Bat. C - projet Magrit Campus Scientifique, 54506 Vandoeuvre-lès-Nancy Cédex, BP 239</wicri:regionArea>
<wicri:noRegion>BP 239</wicri:noRegion>
<wicri:noRegion>BP 239</wicri:noRegion>
</affiliation>
<affiliation wicri:level="1"><country wicri:rule="url">France</country>
</affiliation>
</author>
</analytic>
<monogr></monogr>
<series><title level="s" type="main" xml:lang="en">Lecture Notes in Mathematics</title>
<idno type="ISSN">0075-8434</idno>
<idno type="ISSN">0075-8434</idno>
</series>
</biblStruct>
</sourceDesc>
<seriesStmt><idno type="ISSN">0075-8434</idno>
</seriesStmt>
</fileDesc>
<profileDesc><textClass></textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">In this chapter and in the next one, we describe one of the most popular shape descriptors, Lowe's Scale-Invariant Feature Transform (SIFT) method [114]. In continuation we will perform a structural and practical comparison of the SIFTbased matching method with the Level Line Descriptor method (LLD) developed in this book. The LLD method in fact includes the features of the recent, also popular, MSER method. Comparing SIFT and LLD is not an easy task, since they are of different nature. On the one hand LLD is based on geometrical shape descriptors, rigorously invariant with respect to similarity or affine transformations. Moreover, the method comes with decision rules, either for matching or grouping. On the other hand, SIFT descriptors are local patches which are based on key points and which are just similarity-invariant. The comparison will be based on ad hoc experimental protocols, in the spirit of the SIFT method itself. These protocols check the robustness of local descriptors to all perturbations listed in Sect. 1.2 (Chap. 1). We start with a comprehensive description of the SIFT shape encoding (Sect. 10.1). Then we compare robustness and stability of both shape descriptors (Sect. 10.2). Sect. 10.3 compares the matching performances of both algorithms on pair of images having similar shapes or obtained by photographing the same scene under different viewpoints. The main focus of the book is the computation of matching thresholds. In the SIFT method the thresholds are learned from the pair of images. We shall see that obvious matches can be missed when the query shape appears more than once in the searched image. In the next chapter a fusion of SIFT and of the a contrario techniques both for matching and grouping will be proposed.</div>
</front>
</TEI>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/Main/Merge
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 004407 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Merge/biblio.hfd -nk 004407 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Wicri/Lorraine |area= InforLorV4 |flux= Main |étape= Merge |type= RBID |clé= ISTEX:763F1BE2D6282001100A5C5613EF3840B2713620 |texte= The SIFT Method }}
![]() | This area was generated with Dilib version V0.6.33. | ![]() |