nary op• Data kinds: image × image → image
• Call: this op takes 2 inputs, so fullseye.apply cannot run it (that is the one-image model). g = fullseye.FullseyeGraph(); g.add("out", "convol_image", ["$in1", "$in2"], a=0.5, b=0.5); g.run({"$in1": img1, "$in2": img2}, terminal="out") (names starting with $ are the inputs you pass in)
• HALCON equivalent: convol_image (the HALCON reference is a useful guide to its meaning and parameters)
Convolves with the 2nd input as the filter mask. Corresponds to HALCON's convol_image. ★Like HALCON, it computes a correlation (overlays the mask as written) — scipy's convolve flips the mask, so with an asymmetric mask the image shifts in the opposite direction. The mask may be smaller than the image (the shapes need not match) and is normalized by dividing by the sum of absolute values + 1e-8 (so that a mask summing to 0 does not blow up the division; the output is therefore smaller by a relative 1e-8); borders use reflect. a, b are unused.
• 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.
• gallery_nary_image_algebra — py -3.11 examples/gallery_nary_image_algebra.py
image as input)identity · gaussian · mean_box · bilateral · unsharp · median · min_filter · max_filter
nary)add_image · sub_image · mult_image · div_image · abs_diff_image · max_image · min_image · bit_and
*Provenance: imgops_nary.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.