Matches in SemOpenAlex for { <https://semopenalex.org/work/W58630428> ?p ?o ?g. }
Showing items 1 to 51 of
51
with 100 items per page.
- W58630428 abstract "Edge detection can be formulated as a binary classification problem at the pixel level with the goal of identifying individual pixels as either on-edge or off-edge. To solve this classification problem we use both fixed and adaptive feature selection in conjunction with a support vector machine. This approach provides a direct data-driven solution and does not require the intermediate step of learning a distribution to perform a likelihood-based classification. Furthermore, the approach can readily be adapted for other image processing tasks. The algorithm was tested on a data set of 50 object images, each associated with a hand-drawn 'ground truth' image. We computed ROC curves to evaluate the performance of the general feature extraction and machine learning approach, and compared that to the standard Canny edge detector and with recent work on statistical edge detection. Using a direct pixel-by-pixel error metric enabled us to compare to the statistical edge detection approach, and our algorithm compared favorably. Using a more 'natural' metric enabled comparision with work by the authors of the image data set, and our algorithm performed comparably to the suite of state-of-art edge detectors in that study." @default.
- W58630428 created "2016-06-24" @default.
- W58630428 creator A5037870878 @default.
- W58630428 creator A5049539727 @default.
- W58630428 creator A5075957872 @default.
- W58630428 date "2004-01-01" @default.
- W58630428 modified "2023-09-27" @default.
- W58630428 title "Edge detection in grayscale imagery using machine learning" @default.
- W58630428 hasPublicationYear "2004" @default.
- W58630428 type Work @default.
- W58630428 sameAs 58630428 @default.
- W58630428 citedByCount "0" @default.
- W58630428 crossrefType "proceedings" @default.
- W58630428 hasAuthorship W58630428A5037870878 @default.
- W58630428 hasAuthorship W58630428A5049539727 @default.
- W58630428 hasAuthorship W58630428A5075957872 @default.
- W58630428 hasConcept C115961682 @default.
- W58630428 hasConcept C153180895 @default.
- W58630428 hasConcept C154945302 @default.
- W58630428 hasConcept C162307627 @default.
- W58630428 hasConcept C193536780 @default.
- W58630428 hasConcept C31972630 @default.
- W58630428 hasConcept C41008148 @default.
- W58630428 hasConcept C78201319 @default.
- W58630428 hasConcept C9417928 @default.
- W58630428 hasConceptScore W58630428C115961682 @default.
- W58630428 hasConceptScore W58630428C153180895 @default.
- W58630428 hasConceptScore W58630428C154945302 @default.
- W58630428 hasConceptScore W58630428C162307627 @default.
- W58630428 hasConceptScore W58630428C193536780 @default.
- W58630428 hasConceptScore W58630428C31972630 @default.
- W58630428 hasConceptScore W58630428C41008148 @default.
- W58630428 hasConceptScore W58630428C78201319 @default.
- W58630428 hasConceptScore W58630428C9417928 @default.
- W58630428 hasLocation W586304281 @default.
- W58630428 hasOpenAccess W58630428 @default.
- W58630428 hasPrimaryLocation W586304281 @default.
- W58630428 hasRelatedWork W2048756857 @default.
- W58630428 hasRelatedWork W2077037523 @default.
- W58630428 hasRelatedWork W2181578331 @default.
- W58630428 hasRelatedWork W2269767249 @default.
- W58630428 hasRelatedWork W2347743143 @default.
- W58630428 hasRelatedWork W2393995975 @default.
- W58630428 hasRelatedWork W2412051338 @default.
- W58630428 hasRelatedWork W2763966779 @default.
- W58630428 hasRelatedWork W2950568228 @default.
- W58630428 hasRelatedWork W580442969 @default.
- W58630428 isParatext "false" @default.
- W58630428 isRetracted "false" @default.
- W58630428 magId "58630428" @default.
- W58630428 workType "paratext" @default.