- Rating
- 2
- Confidence
- 3
- Informativeness
- 3
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper investigates the errors in open-vocabulary 3D part segmentation that arise when noisy point features and boundary-mixed responses near part boundaries are directly passed to the point-text matching stage. The paper proposes a method named DGV-SFR, which first performs prompt-conditioned, boundary-aware local denoising, and then leverages a truncated Dirichlet process mixture model and score-driven Wiener dynamics to correct uncertain point states. Final prediction employs the original masked point-text matching rule.
Critical strengths
- The paper targets specific failure modes and provides controlled evidence of performance gains.
- The evaluation covers diverse and useful test conditions while preserving a practical prediction interface.
Critical weaknesses
-
The described inference dynamics fail to substantiate a core part of the conservative correction interpretation. The drift contains an anchor term proportional to , yet each block starts from and uses only one deterministic inference step (Eq. (6); l270-272; Algorithm A2). Consequently, at this step, the anchor term is identically zero. The lower mIoU of the separately trained "w/o anchor tethering" variant (Table 5) does not prove that anchoring directly constrains inference-time motion. Furthermore, Eq. (9) only proves local posterior-density ascent for the continuous-time, score-only trajectory; it does not prove that the full finite-step update preserves part boundaries. This is critical because conservative, boundary-preserving motion is claimed as the primary reason for the method's performance gains.
-
The probabilistic coupling between the mixture posterior and the correction path is not sufficiently specified to justify the claimed unified variational mechanism. The path drift in the generative factor uses a density constructed from the current variational posterior (Eqs. (5)–(7)), while the approximate posterior also includes a path factor. Appendix C.4 appears to describe both path factors as using the same Euler Gaussian transitions. If the variational path factor uses the same transition as the generative factor, the path KL may vanish or become degenerate; if they differ, the respective distributions and the path KL divergence need to be explicitly specified. Moreover, the point-wise uncertainty is computed deterministically from variational statistics rather than being represented as an independent factor in Eq. (7). The ELBO ablation shows that the overall objective matters, but fails to disentangle which part of the claimed joint trajectory inference actually drives the performance improvement. This ambiguity affects my assessment of the paper's primary theoretical contribution.
Questions for authors
-
What are the exact number of training steps and step size ? Does inference indeed execute exactly one Euler step per recovery block as described? On a fixed trained checkpoint, please compare inference results with the anchor term enabled versus disabled, and report the mean mIoU and boundary metrics. If the claimed inference benefit requires more than one step, a small-scale controlled sweep over the number of deterministic inference steps would help clarify the discrepancy with Table 5.
-
Please present the generative and variational path transition densities side by side, clarify whether the generative transition depends on the current variational parameters, and explicitly show the path contribution to the ELBO. Is this contribution non-zero during training? If so, please report an ablation study that removes only the path-specific term while retaining the mixture ELBO, segmentation loss , and boundary supervision .
-
For the described one-step inference update, using the existing test protocol, please report the fraction of corrected points for which the modeled log-density decreases, along with the boundary leakage rate and boundary mIoU. Comparing a fixed checkpoint with and without the boundary projection operator would help disentangle the operator's direct effect on inference-time motion from the effect of retraining the model after removing this term in Table 5.
5 comments
[Q2] Generative and variational path transitions and the ELBO contribution
Conditioned on the same current variational surface and drift, the generative and variational transitions are the same Euler--Gaussian kernel (Sec. 3.3, Eqs. (5)--(8); Appendix C.3--C.4):
The drift f_phi^l depends on the current variational mixture through its posterior surface. The path contribution to the evidence lower bound is zero because these conditional kernels coincide. Let q denote q_phi^l. The four remaining ELBO terms are:
Here V^l denotes the stick-breaking variables, \theta^l the component parameters, C^l the categorical component assignments, and Y^l the observed initial semantic states. Thus, , , and respectively account for the stick, component, and assignment prior--posterior contributions, while is the likelihood of the observed states. The resulting decomposition is:
The final 0 is the Wiener-path contribution, which vanishes because the generative and variational path kernels coincide. The component prior is defined by:
and is independent of the stick variables. The path term is zero because the conditional kernels coincide; segmentation and boundary log-likelihoods are added outside the ELBO (Sec. 3.4, Eqs. (10)--(11); Appendix C.4). Thus there is no nonzero path-only term to ablate. We have clarified the factorization and described the contribution as posterior-surface estimation with constrained Wiener correction.
[Q3] Density decrease, boundary leakage, boundary mIoU, and projection
In the fixed-checkpoint comparison, a point is corrected when its state displacement satisfies:
The boundary band contains points with at least one different-part neighbor among 24 neighbors; the neighborhood size is:
Leakage is the fraction of boundary-band updates directed toward an incompatible part. Density decrease is the fraction of corrected points satisfying:
mIoU averages four prompt/pose settings; boundary metrics use canonical Part-of-Object prompts.
| Benchmark | Projection | Mean mIoU | Boundary mIoU | Leakage (%) | Density decrease (%) |
|---|---|---|---|---|---|
| Coarse | Off | 55.800 | 39.1 | 9.8 | 3.8 |
| Coarse | On | 55.840 | 40.1 | 8.2 | 3.7 |
| Fine | Off | 32.200 | 24.3 | 13.5 | 5.2 |
| Fine | On | 32.295 | 25.0 | 11.7 | 5.1 |
| ShapeNetPart | Off | 37.350 | 32.7 | 14.4 | 7.1 |
| ShapeNetPart | On | 37.460 | 33.8 | 12.1 | 7.0 |
| PartNetE | Off | 21.620 | 31.1 | 15.5 | 9.0 |
| PartNetE | On | 21.703 | 32.3 | 14.0 | 8.8 |
| Macro average | Off | 36.743 | 31.800 | 13.300 | 6.275 |
| Macro average | On | 36.824 | 32.800 | 11.500 | 6.150 |
Projection changes overall mIoU slightly, improves boundary mIoU, and lowers leakage on all four benchmarks. This supports an empirical reduction in risky boundary motion, not a universal boundary-preservation guarantee (Appendix C.5; Appendix F.3).
I have reviewed the rebuttal and additional diagnostics, which resolve my initial technical questions regarding the one-step inference dynamics and step sweep, the vanishing ELBO path term, and the fixed-checkpoint boundary projection and density-decrease evaluations. However, these clarifications—including the zero path KL contribution, the inactive anchor force during default inference, and the newly disclosed supervised training on 3DCoMPaT200 (which is not supervision-matched with Table 1's zero-shot baselines)—substantially narrow the paper's theoretical and empirical claims, while a simpler boundary-aware refinement baseline under matched supervision is still needed. Because properly integrating these major revisions and protocol alignments requires a thorough reorganization beyond the rebuttal period, I maintain my rating of 2 (Weak rejection) and encourage the authors to revise and resubmit.
Dear Reviewer 77EZ,
Thank you for examining the inference dynamics and probabilistic formulation. We answer each weakness and question separately below.
[W1] Anchor displacement and the scope of boundary claims
Each block starts at Yiℓ,0=yiℓ,0 and uses one deterministic inference step, so the anchor force −λ(1−uiℓ)(Yiℓ−yiℓ,0) is zero at that step (Sec. 3.3, Eq. (6); Appendix C.3). The separately trained Table 6 variants (36.82 and 36.22 mIoU) therefore do not demonstrate a direct fixed-checkpoint inference effect. Equation (9) proves local density ascent only for the continuous-time score component; it does not guarantee boundary preservation by the full discrete update (Appendix C.5). The projection diagnostics support only an empirical reduction in boundary leakage under the evaluated protocol.
[W2] Mixture–path coupling and pointwise uncertainty
The variational mixture defines the posterior semantic surface, whose score drives the Wiener correction (Sec. 3.3, Eqs. (4)–(9); Appendix C.2–C.4). Pointwise uncertainty is computed from assignment entropy and posterior covariance, not represented by an independent variational factor. Because the submitted generative and variational path kernels coincide, their explicit log-ratio is zero. The correction ablations (Sec. 4.4, Tables 4–6) support the complete module but do not isolate a path-only effect; we therefore make no claim of a separate Kullback–Leibler contribution for the path.
[Q1] Training/inference steps, fixed-checkpoint anchor comparison, and step sweep
Training uses Ttrain=2 Euler steps with Δt=0.5; the training noise scale is νtrain=0.006. Default inference uses one deterministic Euler update per recovery block, Tinf=1, with νinf=0 (Sec. 4.1; Appendix E.3). Under the default one-step update, fixed-checkpoint outputs with and without the anchor are mathematically identical because the anchor force is zero at initialization. Their metrics therefore equal the default-run values: overall mIoU 36.824, boundary mIoU 32.800, leakage 11.500%, and density-decrease rate 6.150%. This equality follows analytically from Eq. (6), rather than from an independently measured anchor ablation.
The sweep varies only the number of deterministic inference updates per block:
For one, two, and four steps, displacement/boundary mIoU/leakage/density gain are 0.061/32.8/11.5%/+0.46, 0.104/32.5/12.0%/+0.52, and 0.171/31.9/13.4%/+0.57. More steps increase movement and leakage while reducing boundary mIoU; one step is a conservative default, not a unique optimum. We have reported these diagnostics in Appendix F.3.