fullseye

radon_transform — TOMOGRAPHY forward op

• データ種: image2d → sinogram
• 呼び出し: import fullseye as fs; fs.ledger.radon_transform(image, angles_deg=None, n_detectors=None, oversample=1) (実装を直接呼ぶなら import tomography; tomography.radon_transform(image, angles_deg=None, n_detectors=None, oversample=1)、台帳から引くなら opstomography.get("radon_transform"))

使い方

Forward parallel-beam projection: a slice in, a sinogram out.

Rows of the result are projection angles and columns are detector bins. The convention is fixed here and never negotiated again; every other operator in this module reads it the same way, and a transposed sinogram is structurally indistinguishable from a valid one (see :mod:opstomography for the measured consequences of not giving it its own sort).

The ray at detector bin j and angle theta is the line x cos(theta) + y sin(theta) = j - (n_det-1)/2, with x the column offset from the image centre and y the row offset — so +y runs down the array, matching the rest of Fullseye’s image indexing rather than a textbook’s upward y. The transform is the same either way (the sinogram is mirrored in the angle axis), but only one of the two agrees with :func:ellipse_phantom, and the tests hold them together.

Accuracy against the closed form (a disc of radius 60 px in a 256-px grid, 180 views), measured in tests/test_tomography.py: interior RMS error 0.073 % of the peak line integral, whole-sinogram RMS 0.402 % — the difference between the two being the partial-volume edge, where the phantom’s own anti-aliased boundary is what is being sampled.

:param image: 2-D slice, at least 2x2. :param angles_deg: 1-D view angles in degrees; None -> linspace(0, 180, 180, endpoint=False). :param n_detectors: bins; None -> odd count covering the diagonal. :param oversample: ray samples per pixel, 1 .. 8. The default is 1 because 4 measures no better (0.073 % against 0.070 %). :returns: (n_angles, n_detectors) float64 sinogram. :raises ValueError: on non-finite input, an empty angle list, a detector count under 4, or a sinogram over :data:MAX_SINOGRAM_ELEMENTS.

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。
• 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。

実行できる例(この op を実際に呼ぶ検証済みサンプル)

• ct_reconstruction — py -3.11 examples/ct_reconstruction.py
• poc_ct_fidelity — py -3.11 examples/poc_ct_fidelity.py

型が繋がる次の op(sinogram を入力に取れる)

backproject_sinogram · filtered_backprojection · sart_reconstruct · beam_hardening_apply · beam_hardening_correct · ring_artifact_apply · ring_artifact_remove · metal_trace_interpolate

同カテゴリ(forward)

ellipse_phantom · ellipse_sinogram

---

Provenance: tomography.py — TOMOGRAPHY 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.

This site is open source. Improve this page.