Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313172897> ?p ?o ?g. }
- W4313172897 endingPage "543" @default.
- W4313172897 startingPage "526" @default.
- W4313172897 abstract "Detecting tiny objects is one of the main obstacles hindering the development of object detection. The performance of generic object detectors tends to drastically deteriorate on tiny object detection tasks. In this paper, we point out that either box prior in the anchor-based detector or point prior in the anchor-free detector is sub-optimal for tiny objects. Our key observation is that the current anchor-based or anchor-free label assignment paradigms will incur many outlier tiny-sized ground truth samples, leading to detectors imposing less focus on the tiny objects. To this end, we propose a Gaussian Receptive Field based Label Assignment (RFLA) strategy for tiny object detection. Specifically, RFLA first utilizes the prior information that the feature receptive field follows Gaussian distribution. Then, instead of assigning samples with IoU or center sampling strategy, a new Receptive Field Distance (RFD) is proposed to directly measure the similarity between the Gaussian receptive field and ground truth. Considering that the IoU-threshold based and center sampling strategy are skewed to large objects, we further design a Hierarchical Label Assignment (HLA) module based on RFD to achieve balanced learning for tiny objects. Extensive experiments on four datasets demonstrate the effectiveness of the proposed methods. Especially, our approach outperforms the state-of-the-art competitors with 4.0 AP points on the AI-TOD dataset. Codes are available at https://github.com/Chasel-Tsui/mmdet-rfla ." @default.
- W4313172897 created "2023-01-06" @default.
- W4313172897 creator A5057164374 @default.
- W4313172897 creator A5065952768 @default.
- W4313172897 creator A5067222034 @default.
- W4313172897 creator A5069540177 @default.
- W4313172897 creator A5073032922 @default.
- W4313172897 creator A5076179689 @default.
- W4313172897 date "2022-01-01" @default.
- W4313172897 modified "2023-10-17" @default.
- W4313172897 title "RFLA: Gaussian Receptive Field Based Label Assignment for Tiny Object Detection" @default.
- W4313172897 cites W1861492603 @default.
- W4313172897 cites W2037227137 @default.
- W4313172897 cites W2117539524 @default.
- W4313172897 cites W2132481658 @default.
- W4313172897 cites W2194775991 @default.
- W4313172897 cites W2565639579 @default.
- W4313172897 cites W2570343428 @default.
- W4313172897 cites W2625219738 @default.
- W4313172897 cites W2768489488 @default.
- W4313172897 cites W2885641989 @default.
- W4313172897 cites W2896155169 @default.
- W4313172897 cites W2899607431 @default.
- W4313172897 cites W2913012218 @default.
- W4313172897 cites W2962766617 @default.
- W4313172897 cites W2963307811 @default.
- W4313172897 cites W2963351448 @default.
- W4313172897 cites W2963857746 @default.
- W4313172897 cites W2964093967 @default.
- W4313172897 cites W2964241181 @default.
- W4313172897 cites W2982770724 @default.
- W4313172897 cites W2985384565 @default.
- W4313172897 cites W2986357608 @default.
- W4313172897 cites W2988452521 @default.
- W4313172897 cites W2997408160 @default.
- W4313172897 cites W3009396058 @default.
- W4313172897 cites W3011688396 @default.
- W4313172897 cites W3012573144 @default.
- W4313172897 cites W3034971973 @default.
- W4313172897 cites W3035396860 @default.
- W4313172897 cites W3046174881 @default.
- W4313172897 cites W3090598983 @default.
- W4313172897 cites W3096609285 @default.
- W4313172897 cites W3098090606 @default.
- W4313172897 cites W3102206498 @default.
- W4313172897 cites W3102701618 @default.
- W4313172897 cites W3106250896 @default.
- W4313172897 cites W3108849448 @default.
- W4313172897 cites W3120031785 @default.
- W4313172897 cites W3159999130 @default.
- W4313172897 cites W3160733999 @default.
- W4313172897 cites W3167308647 @default.
- W4313172897 cites W3172752666 @default.
- W4313172897 cites W3175630421 @default.
- W4313172897 cites W3176081225 @default.
- W4313172897 cites W3177257155 @default.
- W4313172897 cites W3179888767 @default.
- W4313172897 cites W3201797941 @default.
- W4313172897 cites W4206471589 @default.
- W4313172897 cites W4283717190 @default.
- W4313172897 doi "https://doi.org/10.1007/978-3-031-20077-9_31" @default.
- W4313172897 hasPublicationYear "2022" @default.
- W4313172897 type Work @default.
- W4313172897 citedByCount "14" @default.
- W4313172897 countsByYear W43131728972022 @default.
- W4313172897 countsByYear W43131728972023 @default.
- W4313172897 crossrefType "book-chapter" @default.
- W4313172897 hasAuthorship W4313172897A5057164374 @default.
- W4313172897 hasAuthorship W4313172897A5065952768 @default.
- W4313172897 hasAuthorship W4313172897A5067222034 @default.
- W4313172897 hasAuthorship W4313172897A5069540177 @default.
- W4313172897 hasAuthorship W4313172897A5073032922 @default.
- W4313172897 hasAuthorship W4313172897A5076179689 @default.
- W4313172897 hasBestOaLocation W43131728972 @default.
- W4313172897 hasConcept C103278499 @default.
- W4313172897 hasConcept C115961682 @default.
- W4313172897 hasConcept C121332964 @default.
- W4313172897 hasConcept C146849305 @default.
- W4313172897 hasConcept C153180895 @default.
- W4313172897 hasConcept C154945302 @default.
- W4313172897 hasConcept C163716315 @default.
- W4313172897 hasConcept C19071747 @default.
- W4313172897 hasConcept C202444582 @default.
- W4313172897 hasConcept C2524010 @default.
- W4313172897 hasConcept C2776151529 @default.
- W4313172897 hasConcept C2781238097 @default.
- W4313172897 hasConcept C28719098 @default.
- W4313172897 hasConcept C31972630 @default.
- W4313172897 hasConcept C33923547 @default.
- W4313172897 hasConcept C41008148 @default.
- W4313172897 hasConcept C62520636 @default.
- W4313172897 hasConcept C76155785 @default.
- W4313172897 hasConcept C94915269 @default.
- W4313172897 hasConcept C9652623 @default.
- W4313172897 hasConceptScore W4313172897C103278499 @default.
- W4313172897 hasConceptScore W4313172897C115961682 @default.
- W4313172897 hasConceptScore W4313172897C121332964 @default.
- W4313172897 hasConceptScore W4313172897C146849305 @default.