anomaly_eval op• 資料種類:video × video → table
• 呼叫: import fullseye as fs; fs.ledger.anomaly_region_overlap(predicted, masks, min_overlap=0.5)(要直接呼叫實作,import anomeval; anomeval.anomaly_region_overlap(predicted, masks, min_overlap=0.5);從台帳取用則 opsspc.get("anomaly_region_overlap"))
單一工作點下二值缺陷預測與真值各區域的重疊率(table)。
> 以下的詳細說明為原文 —— 摘要與標題已翻譯。
predicted and masks are binary (n, H, W) (or 2-D). For every 8-connected ground-truth region C: overlap = |predicted ∩ C| / |C|; a region counts as detected when overlap >= min_overlap. Returns overlap (per region, in mask order), pro (their mean = one point of the PRO curve), detected_fraction, hit_any (fraction of regions touched by at least one predicted pixel), fpr (predicted good pixels / good pixels), n_regions, min_overlap.
Raises ValueError on empty / mismatched / non-binary input, no ground-truth region, or min_overlap outside (0, 1].
• 範例資料目錄(下載 URL / 授權) —— 2-D 用 skimage.data(BSD/公有領域)加合成圖,3-D 給出真實資料源(Stanford/PDS 等)的下載 URL。
• 運算子來歷與參考文獻 —— 該運算子族所依據的研究/方法出處。
• 演算法的正典(作者、年份)與用途見上面的族使用指南。
• anomaly_eval_metrics — py -3.11 examples/anomaly_eval_metrics.py
• poc_good_sample_variation_model — py -3.11 examples/poc_good_sample_variation_model.py
• poc_patch_anomaly — py -3.11 examples/poc_patch_anomaly.py
table 作為輸入)msa_anova_table · msa_gauge_rr · msa_bias_linearity · msa_attribute_agreement · gum_standard_uncertainty · gum_propagate · gum_expanded · gum_monte_carlo
anomaly_eval)anomaly_image_auroc · anomaly_pixel_auroc · anomaly_pro · anomaly_pixel_f1_max · anomaly_image_f1_max · anomaly_operating_point
*Provenance: anomeval.py — SPC 運算子登記表。本條目由 tools/opdocs.py md 自動產生(請勿手動編輯)。*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.