Matches in SemOpenAlex for { <https://semopenalex.org/work/W3138977901> ?p ?o ?g. }
- W3138977901 abstract "We propose a Dynamic Scale Training paradigm (abbreviated as DST) to mitigate scale variation challenge in object detection. Previous strategies like image pyramid, multi-scale training, and their variants are aiming at preparing scale-invariant data for model optimization. However, the preparation procedure is unaware of the following optimization process that restricts their capability in handling the scale variation. Instead, in our paradigm, we use feedback information from the optimization process to dynamically guide the data preparation. The proposed method is surprisingly simple yet obtains significant gains (2%+ Average Precision on MS COCO dataset), outperforming previous methods. Experimental results demonstrate the efficacy of our proposed DST method towards scale variation handling. It could also generalize to various backbones, benchmarks, and other challenging downstream tasks like instance segmentation. It does not introduce inference overhead and could serve as a free lunch for general detection configurations. Besides, it also facilitates efficient training due to fast convergence. Code and models are available at github.com/yukang2017/Stitcher." @default.
- W3138977901 created "2021-03-29" @default.
- W3138977901 creator A5027961969 @default.
- W3138977901 creator A5045358378 @default.
- W3138977901 creator A5050058710 @default.
- W3138977901 creator A5052856441 @default.
- W3138977901 creator A5060855982 @default.
- W3138977901 creator A5062788756 @default.
- W3138977901 creator A5083397876 @default.
- W3138977901 creator A5091093335 @default.
- W3138977901 date "2020-04-26" @default.
- W3138977901 modified "2023-09-25" @default.
- W3138977901 title "Dynamic Scale Training for Object Detection" @default.
- W3138977901 cites W1536680647 @default.
- W3138977901 cites W1861492603 @default.
- W3138977901 cites W2037227137 @default.
- W3138977901 cites W2102605133 @default.
- W3138977901 cites W2151103935 @default.
- W3138977901 cites W2161969291 @default.
- W3138977901 cites W2194775991 @default.
- W3138977901 cites W2549139847 @default.
- W3138977901 cites W2565639579 @default.
- W3138977901 cites W2599765304 @default.
- W3138977901 cites W2601564443 @default.
- W3138977901 cites W2765407302 @default.
- W3138977901 cites W2768489488 @default.
- W3138977901 cites W2804047946 @default.
- W3138977901 cites W2908113397 @default.
- W3138977901 cites W2912792574 @default.
- W3138977901 cites W2936599103 @default.
- W3138977901 cites W2949117887 @default.
- W3138977901 cites W2954428350 @default.
- W3138977901 cites W2955904076 @default.
- W3138977901 cites W2963016543 @default.
- W3138977901 cites W2963299996 @default.
- W3138977901 cites W2963351448 @default.
- W3138977901 cites W2963381188 @default.
- W3138977901 cites W2963516811 @default.
- W3138977901 cites W2963857746 @default.
- W3138977901 cites W2964121718 @default.
- W3138977901 cites W2964444661 @default.
- W3138977901 cites W2970575838 @default.
- W3138977901 cites W2982770724 @default.
- W3138977901 cites W2987591514 @default.
- W3138977901 cites W2991089415 @default.
- W3138977901 cites W2991391304 @default.
- W3138977901 cites W3018757597 @default.
- W3138977901 cites W3035395201 @default.
- W3138977901 cites W3035396860 @default.
- W3138977901 cites W3046032854 @default.
- W3138977901 cites W3106250896 @default.
- W3138977901 cites W3107867277 @default.
- W3138977901 cites W3110187642 @default.
- W3138977901 cites W639708223 @default.
- W3138977901 cites W8437397 @default.
- W3138977901 doi "https://doi.org/10.48550/arxiv.2004.12432" @default.
- W3138977901 hasPublicationYear "2020" @default.
- W3138977901 type Work @default.
- W3138977901 sameAs 3138977901 @default.
- W3138977901 citedByCount "2" @default.
- W3138977901 countsByYear W31389779012020 @default.
- W3138977901 countsByYear W31389779012023 @default.
- W3138977901 crossrefType "posted-content" @default.
- W3138977901 hasAuthorship W3138977901A5027961969 @default.
- W3138977901 hasAuthorship W3138977901A5045358378 @default.
- W3138977901 hasAuthorship W3138977901A5050058710 @default.
- W3138977901 hasAuthorship W3138977901A5052856441 @default.
- W3138977901 hasAuthorship W3138977901A5060855982 @default.
- W3138977901 hasAuthorship W3138977901A5062788756 @default.
- W3138977901 hasAuthorship W3138977901A5083397876 @default.
- W3138977901 hasAuthorship W3138977901A5091093335 @default.
- W3138977901 hasBestOaLocation W31389779011 @default.
- W3138977901 hasConcept C111919701 @default.
- W3138977901 hasConcept C119857082 @default.
- W3138977901 hasConcept C121332964 @default.
- W3138977901 hasConcept C124101348 @default.
- W3138977901 hasConcept C142575187 @default.
- W3138977901 hasConcept C153294291 @default.
- W3138977901 hasConcept C154945302 @default.
- W3138977901 hasConcept C162324750 @default.
- W3138977901 hasConcept C177264268 @default.
- W3138977901 hasConcept C199360897 @default.
- W3138977901 hasConcept C2524010 @default.
- W3138977901 hasConcept C2776214188 @default.
- W3138977901 hasConcept C2776760102 @default.
- W3138977901 hasConcept C2777211547 @default.
- W3138977901 hasConcept C2777303404 @default.
- W3138977901 hasConcept C2778334786 @default.
- W3138977901 hasConcept C2778755073 @default.
- W3138977901 hasConcept C2779960059 @default.
- W3138977901 hasConcept C2781238097 @default.
- W3138977901 hasConcept C33923547 @default.
- W3138977901 hasConcept C41008148 @default.
- W3138977901 hasConcept C44870925 @default.
- W3138977901 hasConcept C50522688 @default.
- W3138977901 hasConcept C62520636 @default.
- W3138977901 hasConcept C89600930 @default.
- W3138977901 hasConcept C98045186 @default.
- W3138977901 hasConceptScore W3138977901C111919701 @default.
- W3138977901 hasConceptScore W3138977901C119857082 @default.