filtering op• Data kinds: image → image
• Call: fullseye.apply(img, "tf_gradient_domain_reintegrate", a=0.5, b=0.5) (the 2-D model is one image plus two scalar knobs a,b∈[0,1])
gradient_domain_reintegrate: threshold the gradient, then Poisson-reintegrate.
The forward gradient (gx, gy) is computed, gradient vectors whose magnitude is
below a threshold `t = a * max|grad|` are zeroed (small texture/noise
gradients discarded, strong edges kept), and the image is reconstructed from
the modified gradient field by solving the Poisson equation
`lap f = div(g) with an FFT solver. With a == 0` every gradient is kept
and the original image is recovered (up to a constant); with `a > 0` flat
regions are flattened while edges survive -- an edge-preserving gradient-domain
filter. `b` is unused. Output rescaled to [0,1].
• gallery2d_smoothing_rank family guide
• 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.
• gallery2d_smoothing_rank — py -3.11 examples/gallery2d_smoothing_rank.py
image as input)identity · gaussian · mean_box · bilateral · unsharp · median · min_filter · max_filter
filtering)—
*Provenance: ops.py — 2D operator registry. This per-op note is generated by tools/opdocs.py md (do not hand-edit).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.