tb_mat_pinv — 2D typed op

データ種: matrixmatrix

呼び出し: fullseye.apply(img, "tb_mat_pinv", a=0.5, b=0.5) (2-D は 1 画像 + 2 スカラつまみ a,b∈[0,1] のモデル)

使い方

Moore-Penrose pseudo-inverse via SVD, with the cutoff explicit.

Singular values below `rcond * s_max` are treated as zero — that cutoff

*is* the regularisation, so it is a named, documented parameter here

(default `1e-12`) rather than a hidden library default: raising it

discards noisy directions (stabler, more biased), lowering it keeps them

(exact for well-conditioned *A*, explosive near rank deficiency).

Works for any `(m, n): pinv(A) @ b` is the least-squares solution for

`m > n and the minimum-norm solution for m < n`.

HALCON: no direct operator — HALCON reaches the same result through

`svd_matrix` + reciprocal singular values.

Typed bridge of the math op `mat_pinv into the 2-D evolution registry: the same implementation, called under the op(v, a, b) convention. a drives rcond (default 1e-12); b` is unused.

参考(サンプルデータ・文献)

• 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。

実行できる例(この op を実際に呼ぶ検証済みサンプル)

• (まだありません)

型が繋がる次の op(matrix を入力に取れる)

identity · tb_mat_cond · tb_stat_covariance · tb_stat_correlation

同カテゴリ(typed)

tb_points_to_voxel · tb_estimate_point_normals · tb_iss_keypoints · tb_angle_3points · tb_project_points · tb_render_point_depth · tb_statistical_outlier_removal · tb_radius_outlier_removal


*Provenance: ops.py — 2D 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.