Search

On the Structural Failure of Chamfer Distance in 3D Shape Optimization

Topics
MPM
Chamfer Distance
πŸ‘₯ Authors
Chang-Yong Song, David Hyde
🏒 Venue
arXiv Preprint, 2026 (arXiv:2603.09925)
πŸ“„ Status
arXiv Preprint, 2026 (arXiv:2603.09925)
🧠 Keywords
1 more property

Core Claim

Directly optimizing Chamfer Distance (CD) can produce worse CD values than not optimizing it at all. This is not a metric design problem β€” it is a gradient-structural failure.

Why It Fails (3 Propositions)

β€’
Prop 1. The unique attractor of the forward CD gradient is many-to-one collapse β€” multiple source points converge to the same target point
β€’
Prop 2. The reverse (tβ†’s) term provides nonzero gradient to at most 1 of k collapsed points β†’ the remaining kβˆ’1 are stuck in a zero-gradient deadlock
β€’
Prop 3. Local regularizers (repulsion, smoothness, DCD) cannot alter cluster-level drift β€” translational invariance guarantees pairwise forces cancel at the centroid

The Fix (Corollary 1)

Collapse suppression requires coupling that propagates beyond local neighborhoods.
Method
Global Coupling
Result
DCO
βœ—
collapse
DCO + Repulsion
βœ— (local)
collapse
DCD
βœ— (local reweighting)
collapse
MPM prior
βœ“ (shared Eulerian grid)
suppressed
Fourier basis (2D)
βœ“ (shared parameters)
suppressed

Results

β€’
20 directed shape pairs: our method improves two-sided CD on 16/20
β€’
Dragon (topologically complex): 2.5Γ— improvement over physics-only at 4ppc
β€’
Consistent gains confirmed by Hausdorff distance and F1-score

Practical Guidelines

DCO / Physics Ratio
Risk
Recommendation
≀ 1Γ— (nearly convex)
Low
CD alone is acceptable
1.3–2.0Γ—
Moderate
Global coupling recommended
> 2Γ—
High
CD without global coupling strongly discouraged

One-Line Takeaway

When using CD as a training loss, the architecture must provide non-local coupling. Redesigning the metric alone cannot resolve the failure.