Matches in SemOpenAlex for { <https://semopenalex.org/work/W2900897267> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W2900897267 abstract "In this work, we investigate strategies for training convolutional neural networks (CNNs) to perform recognition on remote sensing imagery. In particular we consider the particular problem of semantic segmentation in which the goal is to obtain a dense pixel-wise labeling of the input imagery. Remote sensing imagery is usually stored in the form of very large images, called “tiles”, which are too big to be segmented directly using most CNNs and their associated hardware. Therefore smaller sub-images, called “patches”, must be extracted from the available tiles. A popular strategy in the literature is to randomly sample patches from the tiles. However, in this work we demonstrate experimentally that extracting patches randomly from a uniform, non-overlapping spatial grid, leads to more accurate models. Our findings suggest the performance improvements are the result of reducing redundancy within the training dataset. We also find that sampling mini-batches of patches (for stochastic gradient descent) using constraints that maximizes the diversity of images within each batch leads to more accurate models. For example, in this work we constrained patches to come from varying tiles, or cities. These simple strategies contributed to our winning entry (in terms of overall performance) in the first year of the INRIA Building Labeling Challenge." @default.
- W2900897267 created "2018-11-29" @default.
- W2900897267 creator A5004578492 @default.
- W2900897267 creator A5007239332 @default.
- W2900897267 creator A5054438743 @default.
- W2900897267 creator A5081855980 @default.
- W2900897267 creator A5084623752 @default.
- W2900897267 date "2018-07-01" @default.
- W2900897267 modified "2023-10-16" @default.
- W2900897267 title "On The Extraction of Training Imagery from Very Large Remote Sensing Datasets for Deep Convolutional Segmenatation Networks" @default.
- W2900897267 cites W1686810756 @default.
- W2900897267 cites W1901129140 @default.
- W2900897267 cites W2031489346 @default.
- W2900897267 cites W2163605009 @default.
- W2900897267 cites W2194775991 @default.
- W2900897267 cites W2469938794 @default.
- W2900897267 cites W2950668883 @default.
- W2900897267 cites W2963881378 @default.
- W2900897267 doi "https://doi.org/10.1109/igarss.2018.8519523" @default.
- W2900897267 hasPublicationYear "2018" @default.
- W2900897267 type Work @default.
- W2900897267 sameAs 2900897267 @default.
- W2900897267 citedByCount "1" @default.
- W2900897267 countsByYear W29008972672020 @default.
- W2900897267 crossrefType "proceedings-article" @default.
- W2900897267 hasAuthorship W2900897267A5004578492 @default.
- W2900897267 hasAuthorship W2900897267A5007239332 @default.
- W2900897267 hasAuthorship W2900897267A5054438743 @default.
- W2900897267 hasAuthorship W2900897267A5081855980 @default.
- W2900897267 hasAuthorship W2900897267A5084623752 @default.
- W2900897267 hasConcept C108583219 @default.
- W2900897267 hasConcept C111919701 @default.
- W2900897267 hasConcept C124504099 @default.
- W2900897267 hasConcept C152124472 @default.
- W2900897267 hasConcept C153180895 @default.
- W2900897267 hasConcept C154945302 @default.
- W2900897267 hasConcept C160633673 @default.
- W2900897267 hasConcept C185592680 @default.
- W2900897267 hasConcept C198531522 @default.
- W2900897267 hasConcept C206688291 @default.
- W2900897267 hasConcept C31972630 @default.
- W2900897267 hasConcept C41008148 @default.
- W2900897267 hasConcept C43617362 @default.
- W2900897267 hasConcept C50644808 @default.
- W2900897267 hasConcept C52622490 @default.
- W2900897267 hasConcept C81363708 @default.
- W2900897267 hasConcept C89600930 @default.
- W2900897267 hasConceptScore W2900897267C108583219 @default.
- W2900897267 hasConceptScore W2900897267C111919701 @default.
- W2900897267 hasConceptScore W2900897267C124504099 @default.
- W2900897267 hasConceptScore W2900897267C152124472 @default.
- W2900897267 hasConceptScore W2900897267C153180895 @default.
- W2900897267 hasConceptScore W2900897267C154945302 @default.
- W2900897267 hasConceptScore W2900897267C160633673 @default.
- W2900897267 hasConceptScore W2900897267C185592680 @default.
- W2900897267 hasConceptScore W2900897267C198531522 @default.
- W2900897267 hasConceptScore W2900897267C206688291 @default.
- W2900897267 hasConceptScore W2900897267C31972630 @default.
- W2900897267 hasConceptScore W2900897267C41008148 @default.
- W2900897267 hasConceptScore W2900897267C43617362 @default.
- W2900897267 hasConceptScore W2900897267C50644808 @default.
- W2900897267 hasConceptScore W2900897267C52622490 @default.
- W2900897267 hasConceptScore W2900897267C81363708 @default.
- W2900897267 hasConceptScore W2900897267C89600930 @default.
- W2900897267 hasLocation W29008972671 @default.
- W2900897267 hasOpenAccess W2900897267 @default.
- W2900897267 hasPrimaryLocation W29008972671 @default.
- W2900897267 hasRelatedWork W2480078828 @default.
- W2900897267 hasRelatedWork W2771077876 @default.
- W2900897267 hasRelatedWork W2793276759 @default.
- W2900897267 hasRelatedWork W2888799854 @default.
- W2900897267 hasRelatedWork W2896796067 @default.
- W2900897267 hasRelatedWork W2903479328 @default.
- W2900897267 hasRelatedWork W2926837263 @default.
- W2900897267 hasRelatedWork W2949436339 @default.
- W2900897267 hasRelatedWork W2958768386 @default.
- W2900897267 hasRelatedWork W2972623730 @default.
- W2900897267 hasRelatedWork W2989888859 @default.
- W2900897267 hasRelatedWork W2991441757 @default.
- W2900897267 hasRelatedWork W2995801068 @default.
- W2900897267 hasRelatedWork W3011932222 @default.
- W2900897267 hasRelatedWork W3100667366 @default.
- W2900897267 hasRelatedWork W3120759590 @default.
- W2900897267 hasRelatedWork W3131754117 @default.
- W2900897267 hasRelatedWork W3179749401 @default.
- W2900897267 hasRelatedWork W3196744391 @default.
- W2900897267 hasRelatedWork W2903387688 @default.
- W2900897267 isParatext "false" @default.
- W2900897267 isRetracted "false" @default.
- W2900897267 magId "2900897267" @default.
- W2900897267 workType "article" @default.