features op• データ種: images → table
• 呼び出し: import fullseye as fs; fs.ledger.feat_image_vector(images=None, features=('mean', 'std', 'p01', 'p99', 'rms_contrast', 'michelson_contrast', 'grad_energy', 'laplacian_var'), data_range=None, unit='DN', roi=None) -> 'dict' (実装を直接呼ぶなら import multivar; multivar.feat_image_vector(images=None, features=('mean', 'std', 'p01', 'p99', 'rms_contrast', 'michelson_contrast', 'grad_energy', 'laplacian_var'), data_range=None, unit='DN', roi=None) -> 'dict'、台帳から引くなら opsmultivar.get("feat_image_vector"))
画像 -> 特徴量の表 {"matrix" (N, F), "names", "units", ...}(画像 1 枚が 1 行)。
> 以下の詳細説明は原文のままです —— 要約と見出しは訳出済み。
The bridge from pictures to the multivariate ops: each image becomes a row of named numbers with units, ready for mv_pca / mv_pls / rsm_fit. Features (see FEATURES): mean, std, p01, p50, p99, rms_contrast, michelson_contrast, grad_energy, laplacian_var need no range; saturated_fraction, dark_fraction, entropy_bits need data_range=(lo, hi) — the range of the *sensor or format* (e.g. (0, 4095) for 12-bit), which the data cannot tell: the brightest pixel of a dim image is not the white level. It is therefore required, never inferred, and pixels outside it are refused (the range given is then wrong).
unit names the grey-level unit ("DN" by default) used in the unit strings. roi=(r0, r1, c0, c1) restricts every feature to that window (same for all images).
• design_of_experiments ファミリ ガイド
• サンプルデータ カタログ(DL URL / ライセンス) — 2-D は skimage.data(BSD/public)+ 合成、3-D は実データ源(Stanford/PDS 等)の DL URL。
• 演算子の来歴・参考文献 — この op 族の元になった研究/手法の出典。
• アルゴリズムの正典(著者・年)と用途は上記ファミリ使い方ガイドに記載。
• multivar_inspection_tour — py -3.11 examples/multivar_inspection_tour.py
table を入力に取れる)—
features)—
*Provenance: multivar.py — MULTIVAR operator registry. この per-op ノートは tools/opdocs.py md が自動生成(手編集しない)。*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.