code wiki / _hdl_build / nx_recon3d_gate.nx
nx_recon3d_gate.nx
buildroot/runtime/_hdl_build/nx_recon3d_gate.nx
about
nx_recon3d_gate.nx -- benchmark the classical multi-view triangulation (cadtwin P1). KAT the fixed-point
(projection<->ray consistency, 2-view midpoint recovers a known point), then the CHAMFER benchmark the DTU
SOTA uses: synthesize 4 camera views of a KNOWN 9-point cloud (a ~160-unit part), triangulate each point,
Chamfer vs ground truth -- CLEAN (near-exact) + NOISY (+/-3px, graceful) + NEG-CONTROL (wrong correspondences
=> large Chamfer, so the metric discriminates). Reports Chamfer in units + permille of the object diagonal.
expect_exit: 0 license_tier: ORIGINAL
dependencies 1 imports · 0 importers
imports: nx_recon3d.nx
imported by: nobody (leaf or entry point)
call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown
structs
| none |
consts
| none |
functions
| 9 | func rg_puts(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 } |
| 10 | func rg_putn(v: i64) -> i64 |
| 23 | func rg_tooth(name: *u8, pass: i64, fails: *i64) -> i64 |
| 31 | func rg_raydist(px: i64, py: i64, pz: i64, cx: i64, cy: i64, cz: i64, dx: i64, dy: i64, dz: i64) -> i64 |
| 42 | func main() -> i64 |