Loading...
Loading...
Contributors:
Dataset URI:
10.5281/zenodo.23354705Supplementary data and code for the manuscript "Cross-view agreement rescoring improves tiled small-object detection without retraining" (Scientific Reports, 2026; version 1.1.0 accompanies the revised manuscript). Contents: (1) the full 110-configuration result grid (detector x benchmark x view generator x merge rule) as CSV/XLSX; (2) per-detection dumps (Parquet) for every configuration, from which all tables and figures are computed; (3) the numbers behind every table and figure; (4) the archived experimental pipeline (view generation, agreement computation, rescoring, evaluation and table generation), released under the AGPL-3.0 licence. Data files are released under CC BY 4.0; the code archive retains its AGPL-3.0 licence. The benchmark images (VisDrone2019-DET, DOTA-v1.5, AI-TOD-v2) are not redistributed and must be obtained from their original distributors. The benchmark images of Argoverse-HD / Argoverse 1.1 are likewise not redistributed.Changes in v1.1.0 (manuscript revision): (1) the reproduction script now passes without PyTorch (the NumPy fallback of class-wise NMS matches torchvision exactly); (2) absolute local paths were removed from manifests; (3) code and JSON outputs of all analyses added in the revision; (4) new detection dumps for the non-aerial Argoverse-HD benchmark and for native oriented boxes on DOTA-v1.5; (5) the wording "sealed" was replaced by a description of the self-imposed held-out test protocol. The two new dump archives are uploaded as byte-split parts (.001, .002, ...) because of the file-count limit of the record; README.md explains how to join them. See README.md. The concept DOI 10.5281/zenodo.22256019 always resolves to the latest version.
Keyword(s):
AI-TOD; cross-view agreement; detection rescoring; DOTA; SAHI.
Grants and Funding:
No Grants or Funding Data Available
Associated Publication(s):
No Associated Publication Data Available
Supplementary Data:
No Supplementary Data Available
Related Dataset(s):
No Related Dataset Data Available
The National Library of Medicine (NLM) provides information about biomedical datasets. Inclusion on an NLM Web page does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: NLM Web Policies (nih.gov)