Matches in SemOpenAlex for { <https://semopenalex.org/work/W2335271330> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W2335271330 endingPage "10" @default.
- W2335271330 startingPage "1" @default.
- W2335271330 abstract "To achieve visual-based automatic product image classification, this paper proposed the dual layer classification with diverse complementary features and presented a comparative study of the ways to fuse multiple SVM classifiers. The base SVM classifiers were trained with different type of features (such as local and globa , as well as appearance, shape and texture) and different spatial levels. The output of the base classifiers could be either class labels or probabilities. Two kinds of combination methods were experimentally studied: the fixed rules and stacking schemes. The former combined the output of the base classifiers with some predefined rules while the latter employed a high-level SVM classifier for combining. Experiments on product dataset PI100 showed that the stacking methods yield much better performance than the fixed rules, and all the probability-based method performed better than the label-based ones. The probability based scheme of the dual layer stacking won the best average classification accuracy of 86.9% on the public product dataset PI 100 which indicated a good direction for implementing automatic product image classification in practice." @default.
- W2335271330 created "2016-06-24" @default.
- W2335271330 creator A5000212361 @default.
- W2335271330 creator A5082846377 @default.
- W2335271330 date "2011-10-31" @default.
- W2335271330 modified "2023-09-27" @default.
- W2335271330 title "Combining Multiple SVM Classifiers For Product Images Classification: A Comparative Study" @default.
- W2335271330 cites W1510526001 @default.
- W2335271330 cites W1542601440 @default.
- W2335271330 cites W1566135517 @default.
- W2335271330 cites W1963799115 @default.
- W2335271330 cites W1970193322 @default.
- W2335271330 cites W1987281309 @default.
- W2335271330 cites W1988790447 @default.
- W2335271330 cites W1999063582 @default.
- W2335271330 cites W2002247107 @default.
- W2335271330 cites W2029458478 @default.
- W2335271330 cites W2030463866 @default.
- W2335271330 cites W2047541160 @default.
- W2335271330 cites W2089481269 @default.
- W2335271330 cites W2125613399 @default.
- W2335271330 cites W2145073242 @default.
- W2335271330 cites W2145607950 @default.
- W2335271330 cites W2153635508 @default.
- W2335271330 cites W2154683974 @default.
- W2335271330 cites W2158275940 @default.
- W2335271330 cites W2158869137 @default.
- W2335271330 cites W2162915993 @default.
- W2335271330 cites W2163352848 @default.
- W2335271330 cites W2186094539 @default.
- W2335271330 cites W2313140711 @default.
- W2335271330 cites W2342270547 @default.
- W2335271330 cites W2912934387 @default.
- W2335271330 cites W91254240 @default.
- W2335271330 doi "https://doi.org/10.4156/jdcta.vol5.issue10.1" @default.
- W2335271330 hasPublicationYear "2011" @default.
- W2335271330 type Work @default.
- W2335271330 sameAs 2335271330 @default.
- W2335271330 citedByCount "1" @default.
- W2335271330 countsByYear W23352713302012 @default.
- W2335271330 crossrefType "journal-article" @default.
- W2335271330 hasAuthorship W2335271330A5000212361 @default.
- W2335271330 hasAuthorship W2335271330A5082846377 @default.
- W2335271330 hasConcept C119857082 @default.
- W2335271330 hasConcept C12267149 @default.
- W2335271330 hasConcept C124101348 @default.
- W2335271330 hasConcept C153180895 @default.
- W2335271330 hasConcept C154945302 @default.
- W2335271330 hasConcept C2524010 @default.
- W2335271330 hasConcept C33923547 @default.
- W2335271330 hasConcept C41008148 @default.
- W2335271330 hasConcept C90673727 @default.
- W2335271330 hasConceptScore W2335271330C119857082 @default.
- W2335271330 hasConceptScore W2335271330C12267149 @default.
- W2335271330 hasConceptScore W2335271330C124101348 @default.
- W2335271330 hasConceptScore W2335271330C153180895 @default.
- W2335271330 hasConceptScore W2335271330C154945302 @default.
- W2335271330 hasConceptScore W2335271330C2524010 @default.
- W2335271330 hasConceptScore W2335271330C33923547 @default.
- W2335271330 hasConceptScore W2335271330C41008148 @default.
- W2335271330 hasConceptScore W2335271330C90673727 @default.
- W2335271330 hasIssue "10" @default.
- W2335271330 hasLocation W23352713301 @default.
- W2335271330 hasOpenAccess W2335271330 @default.
- W2335271330 hasPrimaryLocation W23352713301 @default.
- W2335271330 hasRelatedWork W2041399278 @default.
- W2335271330 hasRelatedWork W2099369243 @default.
- W2335271330 hasRelatedWork W2120008580 @default.
- W2335271330 hasRelatedWork W2136184105 @default.
- W2335271330 hasRelatedWork W2163073107 @default.
- W2335271330 hasRelatedWork W3194539120 @default.
- W2335271330 hasRelatedWork W4205958290 @default.
- W2335271330 hasRelatedWork W4223656335 @default.
- W2335271330 hasRelatedWork W2187500075 @default.
- W2335271330 hasRelatedWork W2345184372 @default.
- W2335271330 hasVolume "5" @default.
- W2335271330 isParatext "false" @default.
- W2335271330 isRetracted "false" @default.
- W2335271330 magId "2335271330" @default.
- W2335271330 workType "article" @default.