gs_render — DRIVE `gsplat` op

• Data kinds: table × matrix → table

• Call: import fullseye as fs; fs.ledger.gs_render(gs: 'dict', pose, K, width: 'int', height: 'int', *, light=(0.3, -0.5, 0.8), ambient: 'float' = 0.35, sky=(0.62, 0.75, 0.92), max_radius: 'int' = 64, max_pairs: 'int' = 12000000, antialias: 'bool' = True) -> 'dict' (to call the implementation directly, import gsplatnp; gsplatnp.gs_render(gs: 'dict', pose, K, width: 'int', height: 'int', *, light=(0.3, -0.5, 0.8), ambient: 'float' = 0.35, sky=(0.62, 0.75, 0.92), max_radius: 'int' = 64, max_pairs: 'int' = 12000000, antialias: 'bool' = True) -> 'dict'; from the registry, opsdrive.get("gs_render"))

Usage

> This operator's description has not been translated yet. The original text follows as it is.

3DGS をカメラ(`pose = world→camera 4×4、K` 3×3、driveworld.world_camera と同じ規約: −Z を見る、行は上が小さい)で描く。

手順: (1) 中心をカメラ系へ、深さ d = −z ≤ 1e-6 は捨てる、(2) 2-D 共分散 Σ' = J W Σ Wᵀ Jᵀ + 0.3 I、

J = [[fx/d, 0, fx·x/d²], [0, −fy/d, −fy·y/d²]]、(3) 半径 r = ⌈3 √λ_max(Σ')⌉(≤ `max_radius`、超える物は捨てずに r を切る)の窓の

画素ごとに α = min(0.99, o · exp(−½ Δᵀ Σ'⁻¹ Δ))(α < 1/255 は捨てる)、(4) 画素ごとに深さの順で手前から

C = Σ c_i α_i T_i、T_i = Π_{j<i}(1 − α_j)、背景 = (1 − Σ α_i T_i) · sky。色は Lambert(world_camera と同じ ambient + (1 − ambient)|n·l|、

光はカメラ系)。すべて配列演算(ガウシアン × 窓の画素の組を平らに並べ、画素 → 深さで並べて群ごとの累積和で T を出す)。

返り値: `color (H,W,3)・alpha (H,W)(= Σ α_i T_i)・depth` (H,W)(α で重みづけた深さ / alpha、alpha < 0.5 は NaN)・

`label (H,W)(重み最大のガウシアンのラベル、alpha < 0.5 は −1)・n_pairs・n_visible`。

fail-closed: 画像の大きさ ≤ 0、max_radius < 1、組の数が `max_pairs` を超える(遅すぎる描画を黙って走らせない)は ValueError。

References (sample data, literature)

• Sample-data catalog (download URLs / licences) — 2-D uses skimage.data (BSD/public domain) plus synthetic images; 3-D lists download URLs for real data sources (Stanford, PDS, …).

• Operator provenance and references — the sources of the research/methods this op family came from.

• The canonical algorithm (author, year) and its uses are named in the family usage guide above.

Runnable examples (verified samples that actually call this op)

• poc_kendama — py -3.11 examples/poc_kendama.py

Ops the type connects to (they accept `table` as input)

course_layout · course_occupancy · course_contains · world_build · world_camera · world_move · lidar_scan · rss_longitudinal_same

Same category (`gsplat`)

gs_from_world · gs_update · gs_render_fn · gs_read_file

*Provenance: gsplatnp.py — DRIVE operator registry. This per-op note is generated by tools/opdocs.py md (do not hand-edit).*

gs_render — DRIVE gsplat op