- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper proposes FreePart, a framework that improves part-level 3D generation by refining how an object is divided into parts. It introduces Part-level Decomposition Assessment using surface complexity and occlusion-based risks, and trains a VLM through SFT and GDPO. The Diagnoser, Strategist, and Evaluator then guide iterative adjacent-part merges through a DAG-based search, accepting or rejecting proposals based on predicted SC/OCC risks and the preservation of part-level logic. It consistently improves part-level and overall 3D generation quality across diverse generation pipelines. The motivation is convincing, but some implementations are not well explained.
Critical strengths
-
The paper treats decomposition as an optimizable component. It offers a well-motivated way to improve 3D generation by refining part decomposition without modifying the generator, making it applicable to existing pipelines.
-
Reported improvements across three generators support the practical value of decomposition refinement.
Critical weaknesses
-
The non-monotonic relationship between OCC and the generation quality, as well as the reference value of 0.6, directly determine the training labels and search decisions, but are insufficiently justified. The paper does not adequately explain how this relationship was established or what evidence supports it, and does not explain why lower occlusion can yield higher errors.
-
The paper does not report end-to-end latency, average VLM call counts, or actual search overhead. FreePart is designed to avoid the cost of exhaustive generation and Sec. 1 emphasizes efficient exploration. Given the sequential execution of three roles using a 72B VLM, it remains unclear whether FreePart achieves better efficiency.
-
The paper states that maps continuous SC values to low, medium, and high risk (Sec. 3), but does not specify the intervals or selection procedure. Their justification and influence on the reported results remain insufficiently examined.
-
The fixed-rule strategy ablation in Table 5 is insufficiently described. Table 5 replaces the adaptive Strategist with fixed rules, but does not specify how these rules select merge candidates.
Questions for authors
-
The paper emphasizes DAG-based search and the complementary roles of the three agents. Table 5 shows that removing a certain capability reduces performance, but it does not isolate the benefits of separating decisions into three roles or adopting the proposed search structure. How would the method perform with more intuitive alternative strategies, such as greedy merging using the same SC/OCC signals, single-agent decision-making, or directly predicting the final grouping? These comparisons may help clarify the benefits of the proposed design.
-
What is the average number of VLM calls per input, the computational cost of the search stage, and the end-to-end runtime?
4 comments
FdZz — Response 2/3
Q3. SC Discretization and Threshold Sensitivity.
The paper states that (Q_{\mathrm{SC}}) maps continuous SC values to low, medium, and high risk (Sec. 3), but does not specify the intervals or selection procedure. Their justification and influence on the reported results remain insufficiently examined.
SC risks are defined using training-set percentile thresholds: low ≤ Q30, medium (Q30, Q70], and high > Q70. We further evaluate three percentile configurations (25/75, 30/70, and 35/65), with 30/70 achieving the highest SC risk classification accuracy, as shown in Table R4.
R4. Sensitivity of SC Risk Prediction
| SC Percentile Cutoffs | SC Thresholds | Classification Accuracy ↑ |
|---|---|---|
| 25% / 75% | 0.00901 / 0.06124 | 82.40% |
| 30% / 70% (Ours) | 0.01667 / 0.05589 | 85.11% |
| 35% / 65% | 0.02338 / 0.05113 | 77.16% |
Separately, risk-stratified reconstruction results show that the high-SC group is more challenging (baseline Part CD 34.51 versus 26.85/26.93 in the low/medium groups) and benefits from FreePart (Part F1-0.1: 36.45 → 44.85).
R5. Reconstruction Performance Across SC Risk Levels.
| SC Risk | Method | Part-level Bbox IoU ↑ | Part-level CD ↓ | Part-level F1-0.1 ↑ | Part-level F1-0.05 ↑ | Overall-level CD ↓ | Overall-level F1-0.1 ↑ | Overall-level F1-0.05 ↑ |
|---|---|---|---|---|---|---|---|---|
| Low | OmniPart | 23.86 | 26.85 | 49.09 | 27.41 | 14.87 | 77.08 | 49.90 |
| Low | OmniPart + FreePart | 26.90 | 23.63 | 55.17 | 32.19 | 13.92 | 80.51 | 54.92 |
| Medium | OmniPart | 23.31 | 26.93 | 49.25 | 27.43 | 14.66 | 77.40 | 52.29 |
| Medium | OmniPart + FreePart | 24.40 | 24.48 | 53.95 | 31.27 | 14.00 | 79.86 | 54.59 |
| High | OmniPart | 20.91 | 34.51 | 36.45 | 19.32 | 16.51 | 76.95 | 60.84 |
| High | OmniPart + FreePart | 24.75 | 31.59 | 44.85 | 23.84 | 16.15 | 77.14 | 61.29 |
| All | OmniPart | 23.31 | 27.37 | 48.44 | 26.94 | 14.80 | 77.30 | 52.21 |
| All | OmniPart + FreePart | 25.07 | 24.69 | 53.72 | 31.06 | 14.13 | 79.85 | 55.05 |
Q4. Fixed-rule Strategist baseline.
The fixed-rule strategy ablation in Table 5 is insufficiently described. Table 5 replaces the adaptive Strategist with fixed rules, but does not specify how these rules select merge candidates.
We apologize for the insufficient description of the fixed-rule Strategist ablation. The Diagnoser and Evaluator remain unchanged, while Strategist is replaced by deterministic contact-based selection. From the current 2D part masks, we enumerate directly contacting and allowed nearby pairs; candidates are ranked by direct contact first, then contact-pixel count, and then proximity-overlap score. The highest-ranked candidate not previously rejected is previewed and passed to the original Evaluator. This ablation changes candidate-selection policy while retaining the remaining diagnosis and verification mechanisms.
FdZz — Response 3/3
Q5. Alternative decision organizations and search.
The paper emphasizes DAG-based search and the complementary roles of the three agents. Table 5 shows that removing a certain capability reduces performance, but it does not isolate the benefits of separating decisions into three roles or adopting the proposed search structure. How would the method perform with more intuitive alternative strategies, such as greedy merging using the same SC/OCC signals, single-agent decision-making, or directly predicting the final grouping? These comparisons may help clarify the benefits of the proposed design.
We add risk-greedy, single-agent iterative, single-agent direct grouping, and three-role direct grouping baselines. The iterative single-agent combines proposal and verification; the two direct-grouping variants predict the final grouping without the same stepwise accepted-merge trajectory. Results from the exploratory comparison subset are:
R6. Comparison of Alternative Search Strategies.
| Method | Bbox IoU ↑ | Part CD ↓ | Part F1-.1 ↑ | Part F1-.05 ↑ | Overall CD ↓ | Overall F1-.1 ↑ | Overall F1-.05 ↑ |
|---|---|---|---|---|---|---|---|
| FreePart | 0.205175 | 0.239538 | 0.559285 | 0.327115 | 0.131650 | 0.814259 | 0.574048 |
| Risk-greedy | 0.203003 | 0.240071 | 0.558539 | 0.325833 | 0.132064 | 0.813540 | 0.572634 |
| Single-agent iterative | 0.196134 | 0.244501 | 0.545202 | 0.317609 | 0.132611 | 0.814236 | 0.571939 |
| Single-agent direct | 0.193600 | 0.244844 | 0.540510 | 0.306785 | 0.132381 | 0.808848 | 0.555346 |
| Three-role direct | 0.195870 | 0.250097 | 0.529476 | 0.297212 | 0.131774 | 0.806470 | 0.570741 |
FreePart obtains the best mean values across the seven reported metrics. In a separate equal-maximum-call comparison, FreePart also improves Overall CD (0.10550 vs. 0.10777) with fewer actual VLM calls (3.9 vs. 5.7) than risk-greedy.
Finally, the DAG is the space of reachable decompositions; the implemented controller follows a constrained sequential path through accepted merges, rather than exhaustive DAG traversal. Rollback rejects an uncommitted candidate preview at the current state and does not undo a previously accepted merge.
Q6. Inference Cost and Runtime.
What is the average number of VLM calls per input, the computational cost of the search stage, and the end-to-end runtime?
As detailed in Q2.1–Q2.2 (Tables R2–R3), FreePart requires an average of 3.9 VLM calls per object, with 260.6 s search time and 328.8 s end-to-end runtime. These measurements provide a breakdown of FreePart's computational overhead and practical inference cost.
Thank you again for your insightful suggestions. We have provided additional experiments on SC/OCC calibration, inference efficiency, threshold sensitivity, and alternative search strategies. We sincerely appreciate your time and effort and would be grateful for any further feedback before the discussion period concludes.
FdZz — Response 1/3
We thank the reviewer for recognizing the motivation, plug-and-play applicability, and effectiveness across different generation pipelines of our work. Below, we address your concerns point by point.
Q1. Non-monotonic OCC trend and empirical reference.
We thank the reviewer for requesting a more explicit basis for the OCC reference. We divide 1,913 objects into five equal-frequency bins according to GT object-level OCC, yielding:
R1. Relationship Between OCC and Reconstruction Quality.
The interval centered near 0.6 has the lowest mean error and highest mean F-scores, providing a data-driven reference for this evaluated setting. We carried out 2,000 bootstrap resamples per bin to examine uncertainty.
Based on our observations of the dataset, parts with low OCC often exhibit thin, sheet-like geometries. Although these parts are largely visible, their thickness is difficult to infer accurately from 2D observations, often leading to geometric distortions and higher reconstruction errors. This provides a plausible explanation for why lower OCC does not necessarily correspond to better generation quality.
Q2. Computational Efficiency and End-to-End Inference Cost.
Q2.1 Actual VLM Calls, Search Overhead, and End-to-End Time.
Candidate previews do not repeatedly invoke the downstream full 3D generator. We benchmark all methods after warm-up, using identical initial masks and generation seed 42 on three A800 GPUs. All three search methods share a maximum budget of six VLM calls per object. The results are summarized below.
R2. Reconstruction Quality and End-to-End Inference Efficiency.
As shown in Table R2, FreePart achieves the best reconstruction quality across all metrics, while requiring 31.6% fewer VLM calls and 18.5% shorter search time than risk-greedy. Although single-path greedy is faster, its reconstruction quality is substantially lower than FreePart's. Compared with search-free OmniPart, FreePart introduces additional computational overhead but significantly improves reconstruction quality. These results demonstrate that FreePart achieves a favorable balance between reconstruction quality and search efficiency.
Q2.2 Stage-Wise Inference Cost and Search Efficiency.
FreePart costs 252.7 s more than search-free OmniPart but completes 57.4 s faster than risk-greedy and saves 31.6% actual VLM calls. Both search variants reduce mean part count from 7.8 to 6.9 on this subset. We therefore distinguish avoiding repeated full generation during candidate search from reducing end-to-end runtime relative to the original generator.
R3. Stage-Wise Breakdown of Inference Time.