typed op• データ種: qimage → qimage
• 呼び出し: fullseye.apply(img, "tb_iqft2", 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)
*図なし: この op は qimage を入力に取る。画像から始まる Studio のプログラムでは型が届かないので、下の「実行できる例」で使い方を見ること。*
中心化されたスペクトルに対する四元数逆フーリエ変換。→ (H, W, 4)。
> 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 exact inverse of :func:qft2 for the same side and the same mu:
measured round-trip error `2.22e-15` for both sides on a standard-normal
`(32, 32, 4) field. The kernel is exp(+mu * 2*pi*(...))` applied on the
side named, and the `1/(H*W)` normalisation is carried here, as in
`numpy.fft.ifft2`.
Using the wrong side does not raise. `iqft2(qft2(q, "left"), "right")`
returns a finite, plausible quaternion image that is simply not `q`:
measured `max|err| = 1.113` on a random colour image whose own range is
`0.9994 (another seed: 1.063 against 1.0), and — the dangerous case — only 0.054` against a range of
`1.076` on a grey-axis-dominated one, which is small enough to survive a
look at the picture. The `side` argument is required at both ends for
exactly this reason, and the two calls must agree: nothing in the data
records which transform produced it, so nothing downstream can catch the
mismatch for you.
Raises `ValueError: *spectrum* is not a valid (H, W, 4)` field;
*side* is not `'left' / 'right'`; *mu* is not a finite non-zero
3-vector.
Typed bridge of the quat op `iqft2 into the 2-D evolution registry: the same implementation, called under the op(v, a, b) convention. This op has no tunable parameter; a and b` are unused.
• 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。
• (まだありません)
qimage を入力に取れる)identity · tb_quaternion_to_rgb · tb_quat_norm · tb_quat_conjugate_image · tb_quat_normalize_image · tb_monogenic_amplitude · tb_monogenic_phase · tb_monogenic_orientation
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.