typed op• 資料種類:qimage → qimage
• 呼叫:fullseye.apply(img, "tb_iqft2", a=0.5, b=0.5)(2-D 的模型是一張圖 + 兩個純量旋鈕 a,b∈[0,1])
*無圖: 該運算子以 qimage 為輸入。從影像開始的 Studio 程式無法到達該型別，請參見下方的可執行範例。*
對置中頻譜做四元數逆傅立葉變換。→ (H, W, 4)。
> 以下的詳細說明為原文 —— 摘要與標題已翻譯。
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.
• 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。
• (尚無)
qimage 作為輸入)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 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.