typed oplightfield → imagefullseye.apply(img, "tb_lf_refocus", 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)
図は合成の入力 128×128 で実際に走らせた出力。左が入力、右が出力。点群は上から見た散布(明るさ = z)、1-D 列は折れ線、体積は z 方向の最大値投影、動画は中央フレーム、複素画像は振幅、絵にならない返り値は値そのもの。
つまみ a は出力を変えない(実測: 0.1 / 0.5 / 0.9 で同一)。
つまみ b は出力を変えない(実測: 0.1 / 0.5 / 0.9 で同一)。
段階(前置きの op → この op。左から順):

別の画像でも(合成シーン / 写真 / 硬貨。上段が入力、下段がその出力。つまみは既定):

Shift-and-add refocus: the synthetic-aperture image focused at slope.
Every view ``(v, u)`` is shifted by ``(-s*(v - v_c), -s*(u - u_c))`` — the
**minus** undoes the parallax of a point at slope ``s`` — and the shifted
views are averaged. Points at that slope add coherently and stay sharp;
everything else is smeared by an amount proportional to its slope
difference times the angular baseline. ``slope=0`` is the plane the array
was already focused on and returns the plain average of the views.
Ground truth it reproduces exactly (pinned in ``tests/test_lightfield.py``):
a single-layer field synthesised at slope ``s0`` and refocused at ``s0``
with ``edge="wrap"`` and an integer ``s0`` returns the original texture to
5.6e-16; sweeping the slope, the variance of the result peaks at ``s0``
(measured exactly on the sweep grid in all 18 texture/slope combinations
listed in the module docstring), and refocusing at ``-s0`` does *not* —
which is the check that catches a flipped shift sign.
Returns a ``(H, W)`` 2-D image.
**Raises** ``ValueError``: *lf* not a valid light field, a non-finite or
over-large *slope* (:data:`MAX_ABS_SLOPE`), unknown *interp* / *edge*.
Typed bridge of the lightfield op lf_refocus 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.
下のプログラムは実際に走ることを確かめてある(図と同じ入力)。Studio のヘルプではこのブロックがボタンになり、その場で読み込んで実行できる。
img_to_lightfield 0.50 0.50
tb_lf_refocus 0.50 0.50
次の例は元の台帳 op lf_refocus を呼ぶもの。この橋渡し op は同じ実装を fn(v, a, b) 規約に合わせただけなので、挙動はそのまま当てはまる(呼び出し形だけ違う)。
py -3.11 examples/lightfield_depth.pypy -3.11 examples/poc_lightfield_depth.pyimage を入力に取れる)identity · gaussian · mean_box · bilateral · unsharp · median · min_filter · max_filter
typed)tb_points_to_voxel · tb_estimate_point_normals · tb_iss_keypoints · tb_project_points · tb_render_point_depth · tb_statistical_outlier_removal · tb_radius_outlier_removal · tb_voxel_grid_downsample
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.