Matches in SemOpenAlex for { <https://semopenalex.org/work/W2956586819> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W2956586819 endingPage "9533" @default.
- W2956586819 startingPage "9524" @default.
- W2956586819 abstract "We present an Active Learning (AL) strategy for re-using a deep Convolutional Neural Network (CNN)-based object detector on a new dataset. This is of particular interest for wildlife conservation: given a set of images acquired with an Unmanned Aerial Vehicle (UAV) and manually labeled gound truth, our goal is to train an animal detector that can be re-used for repeated acquisitions, e.g. in follow-up years. Domain shifts between datasets typically prevent such a direct model application. We thus propose to bridge this gap using AL and introduce a new criterion called Transfer Sampling (TS). TS uses Optimal Transport to find corresponding regions between the source and the target datasets in the space of CNN activations. The CNN scores in the source dataset are used to rank the samples according to their likelihood of being animals, and this ranking is transferred to the target dataset. Unlike conventional AL criteria that exploit model uncertainty, TS focuses on very confident samples, thus allowing a quick retrieval of true positives in the target dataset, where positives are typically extremely rare and difficult to find by visual inspection. We extend TS with a new window cropping strategy that further accelerates sample retrieval. Our experiments show that with both strategies combined, less than half a percent of oracle-provided labels are enough to find almost 80% of the animals in challenging sets of UAV images, beating all baselines by a margin." @default.
- W2956586819 created "2019-07-23" @default.
- W2956586819 creator A5005192117 @default.
- W2956586819 creator A5055174808 @default.
- W2956586819 creator A5071561639 @default.
- W2956586819 creator A5087677326 @default.
- W2956586819 date "2019-12-01" @default.
- W2956586819 modified "2023-10-03" @default.
- W2956586819 title "Half a Percent of Labels is Enough: Efficient Animal Detection in UAV Imagery Using Deep CNNs and Active Learning" @default.
- W2956586819 cites W1594039573 @default.
- W2956586819 cites W1993615002 @default.
- W2956586819 cites W2012256656 @default.
- W2956586819 cites W2046030581 @default.
- W2956586819 cites W2051785999 @default.
- W2956586819 cites W2080211394 @default.
- W2956586819 cites W2107131609 @default.
- W2956586819 cites W2117539524 @default.
- W2956586819 cites W2143897835 @default.
- W2956586819 cites W2194775991 @default.
- W2956586819 cites W2297338375 @default.
- W2956586819 cites W2570343428 @default.
- W2956586819 cites W2737340643 @default.
- W2956586819 cites W2752508182 @default.
- W2956586819 cites W2783822844 @default.
- W2956586819 cites W2810030371 @default.
- W2956586819 cites W2919115771 @default.
- W2956586819 cites W2963037989 @default.
- W2956586819 cites W4206723194 @default.
- W2956586819 cites W4239510810 @default.
- W2956586819 cites W4288076010 @default.
- W2956586819 doi "https://doi.org/10.1109/tgrs.2019.2927393" @default.
- W2956586819 hasPublicationYear "2019" @default.
- W2956586819 type Work @default.
- W2956586819 sameAs 2956586819 @default.
- W2956586819 citedByCount "79" @default.
- W2956586819 countsByYear W29565868192019 @default.
- W2956586819 countsByYear W29565868192020 @default.
- W2956586819 countsByYear W29565868192021 @default.
- W2956586819 countsByYear W29565868192022 @default.
- W2956586819 countsByYear W29565868192023 @default.
- W2956586819 crossrefType "journal-article" @default.
- W2956586819 hasAuthorship W2956586819A5005192117 @default.
- W2956586819 hasAuthorship W2956586819A5055174808 @default.
- W2956586819 hasAuthorship W2956586819A5071561639 @default.
- W2956586819 hasAuthorship W2956586819A5087677326 @default.
- W2956586819 hasBestOaLocation W29565868192 @default.
- W2956586819 hasConcept C108583219 @default.
- W2956586819 hasConcept C119857082 @default.
- W2956586819 hasConcept C140779682 @default.
- W2956586819 hasConcept C146849305 @default.
- W2956586819 hasConcept C150899416 @default.
- W2956586819 hasConcept C153180895 @default.
- W2956586819 hasConcept C154945302 @default.
- W2956586819 hasConcept C189430467 @default.
- W2956586819 hasConcept C2776151529 @default.
- W2956586819 hasConcept C41008148 @default.
- W2956586819 hasConcept C64869954 @default.
- W2956586819 hasConcept C76155785 @default.
- W2956586819 hasConcept C774472 @default.
- W2956586819 hasConcept C81363708 @default.
- W2956586819 hasConcept C94915269 @default.
- W2956586819 hasConceptScore W2956586819C108583219 @default.
- W2956586819 hasConceptScore W2956586819C119857082 @default.
- W2956586819 hasConceptScore W2956586819C140779682 @default.
- W2956586819 hasConceptScore W2956586819C146849305 @default.
- W2956586819 hasConceptScore W2956586819C150899416 @default.
- W2956586819 hasConceptScore W2956586819C153180895 @default.
- W2956586819 hasConceptScore W2956586819C154945302 @default.
- W2956586819 hasConceptScore W2956586819C189430467 @default.
- W2956586819 hasConceptScore W2956586819C2776151529 @default.
- W2956586819 hasConceptScore W2956586819C41008148 @default.
- W2956586819 hasConceptScore W2956586819C64869954 @default.
- W2956586819 hasConceptScore W2956586819C76155785 @default.
- W2956586819 hasConceptScore W2956586819C774472 @default.
- W2956586819 hasConceptScore W2956586819C81363708 @default.
- W2956586819 hasConceptScore W2956586819C94915269 @default.
- W2956586819 hasFunder F4320320924 @default.
- W2956586819 hasIssue "12" @default.
- W2956586819 hasLocation W29565868191 @default.
- W2956586819 hasLocation W29565868192 @default.
- W2956586819 hasOpenAccess W2956586819 @default.
- W2956586819 hasPrimaryLocation W29565868191 @default.
- W2956586819 hasRelatedWork W2296151615 @default.
- W2956586819 hasRelatedWork W2738221750 @default.
- W2956586819 hasRelatedWork W2774444957 @default.
- W2956586819 hasRelatedWork W2896409332 @default.
- W2956586819 hasRelatedWork W2913302899 @default.
- W2956586819 hasRelatedWork W2914010220 @default.
- W2956586819 hasRelatedWork W3021430260 @default.
- W2956586819 hasRelatedWork W3173182854 @default.
- W2956586819 hasRelatedWork W3207637583 @default.
- W2956586819 hasRelatedWork W4225113219 @default.
- W2956586819 hasVolume "57" @default.
- W2956586819 isParatext "false" @default.
- W2956586819 isRetracted "false" @default.
- W2956586819 magId "2956586819" @default.
- W2956586819 workType "article" @default.