- Rating
- 2
- Confidence
- 4
- Informativeness
- 1
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper proposes SAComm, combining sparse patch communication, uncertainty-aware pose correction, and reliability-guided fusion for collaborative perception. Experiments on three datasets demonstrate competitive detection accuracy under pose noise with substantially reduced communication.
Critical strengths
-
The integration of sparse feature transmission and geometric alignment is meaningful. Table 1 demonstrates strong detection performance under pose perturbations while requiring substantially less communication than competing feature-level methods.
-
The component ablations and pose diagnostics in Tables 2-4 provide useful evidence for the effectiveness of pose correction and local fusion, rather than relying solely on end-to-end detection results.
Critical weaknesses
-
Table 1 compares methods at substantially different communication rates, making it difficult to attribute performance gains to the proposed architecture rather than bandwidth differences. Moreover, the methodological distinction from ERMVP and CPPC remains unclear, and CPPC is absent from the main comparison. These issues weaken both the empirical justification and the claimed novelty of SAComm.
-
The proposed ACG module degrades performance under strong pose noise (Tables 2 and 4), while improvements on real-world DAIR-V2X-C become marginal at higher noise levels. Furthermore, evaluation relies primarily on synthetic Gaussian perturbations, with limited statistical validation. These findings raise concerns about whether the claimed robustness generalizes to realistic localization errors.
Questions for authors
-
Can the authors provide communication-rate-matched comparisons with ERMVP, InfoCom, and Where2comm, and explain the absence of CPPC from Table 1? What measurable benefits does patch-based uncertainty modeling provide beyond ERMVP and CPPC?
-
Why does ACG reduce AP70 from 88.12 to 86.98 and increase residual translation error from 0.369 m to 0.545 m at ? Under what conditions does ACG provide consistent improvements?
-
Can the authors substantiate robustness under temporally correlated or biased localization errors and report variability across random seeds or noise realizations, particularly for the marginal improvements on DAIR-V2X-C?
6 comments
Response to Reviewer KcFB [2/4]
W1 / Q1 (Part 2/2): CPPC comparison and uncertainty benefits
Method differences and contribution. ERMVP uses consensus matching, geometric verification, and error modulation on reconstructed features. CPPC optimizes a pose graph from point-cluster correspondences. SAComm uses compressed local patches to retain matching context, preserves candidate displacement distributions, and combines their means and covariances to estimate one correction per sender. Matching statistics and the applied correction are then reused to predict residual pose error and guide local fusion. Our contribution is this joint design of the sparse message and its receiver-side use for alignment and detection at very low communication rates.
CPPC comparison. Table 1 uses a common PointPillars [1] architecture to compare communication and fusion methods while limiting detector differences. CPPC was not included because it builds its sparse detector on FSD [2]. We reproduced CPPC and evaluated its full and low-rate configurations, using descriptor reduction and point sampling for the latter.
Table 2. CPPC comparison: BEV AP70 (%) ↑. Rates are Mbps per sender at 10 Hz. 8D is separately trained; 1/128 is the geometric-point sampling ratio. 128D payloads are estimated from their full fields; 8D rates use encoded byte counts.
| Method | OPV2V | V2XSet | DAIR-V2X-C | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mbps ↓ | 0.2 | 0.4 | 0.6 | Mbps ↓ | 0.2 | 0.4 | 0.6 | Mbps ↓ | 0.2 | 0.4 | 0.6 | ||||
| CPPC 128D, full | 8.022 | 91.76 | 90.52 | 87.49 | 80.97 | 9.008 | 87.85 | 86.75 | 83.93 | 76.01 | 13.001 | 63.13 | 62.77 | 62.12 | 61.74 |
| CPPC 8D, 1/128 | 0.574 | 88.28 | 87.37 | 84.01 | 78.84 | 0.767 | 79.61 | 78.88 | 76.16 | 71.31 | 1.456 | 59.08 | 58.91 | 58.84 | 58.87 |
| SAComm | 0.393 | 92.41 | 91.25 | 89.76 | 86.98 | 0.391 | 82.01 | 79.98 | 77.70 | 75.26 | 0.566 | 65.17 | 61.15 | 58.64 | 57.79 |
Compared with low-rate CPPC, SAComm reduces communication by approximately 31.7%, 49.1%, and 61.1% on OPV2V, V2XSet, and DAIR-V2X-C, respectively, while achieving higher AP70 in 10 of the 12 evaluated settings. On OPV2V and V2XSet, SAComm improves AP70 at all four noise levels, with gains of 8.14 and 3.95 points at . On DAIR-V2X-C, SAComm improves AP70 at ; at , AP70 is 0.20/1.08 points lower while communication is reduced by 61.1%.
Compared with full CPPC, SAComm uses only about 1/20–1/23 of the communication and achieves higher AP70 at every tested noise level on OPV2V. Full CPPC achieves higher AP70 on V2XSet and at nonzero noise on DAIR-V2X-C, but requires substantially more communication. These results highlight SAComm's strong detection performance at very low communication rates.
Evidence for the joint design. In Table 3 of the original submission, at approximately 0.394 Mbps and , individual features with pointwise fusion achieve 84.37% AP70, compared with 86.98% for patches with local fusion. Keeping the patch settings fixed but restricting fusion to reduces AP70 to 83.50%, highlighting the importance of receiver-side use of the retained context. Separately, with OPV2V model weights fixed, we replace per-match PC fitting precision matrices and residual metrics with training-subset means, retaining residual reweighting and the remaining pipeline. Table 3 below shows that this increases translation and yaw errors after PC+ACG at every tested nonzero noise level and reduces AP70 by 0.62–2.22 percentage points. Together, these results support the tested message–receiver design and the benefit of adaptive uncertainty weighting under pose noise.
Table 3. PC uncertainty weighting ablation. BEV AP70 (%); pose errors are measured after PC+ACG, with input pose as the shared reference. Both settings use the same checkpoint and evaluation inputs. Differences use unrounded values. Bold compares post-correction settings.
| Setting | BEV AP70 (%) ↑ | Translation error (m) ↓ | Yaw error (deg) ↓ | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Noise | 0 | 0.2 | 0.4 | 0.6 | 0 | 0.2 | 0.4 | 0.6 | 0 | 0.2 | 0.4 | 0.6 |
| Input pose (reference) | — | — | — | — | 0.0000 | 0.3703 | 0.7426 | 1.0933 | 0.0000 | 0.2227 | 0.4514 | 0.6969 |
| Full model (uncertainty weighting) | 92.41 | 91.25 | 89.76 | 86.98 | 0.0247 | 0.1786 | 0.3023 | 0.5451 | 0.0239 | 0.1560 | 0.2834 | 0.4936 |
| PC mean replacement | 92.52 | 90.63 | 88.06 | 84.77 | 0.0204 | 0.2144 | 0.3843 | 0.6565 | 0.0193 | 0.1793 | 0.3442 | 0.5722 |
References
[1] Lang et al. PointPillars: Fast Encoders for Object Detection From Point Clouds. CVPR 2019.
[2] Fan et al. Fully Sparse 3D Object Detection. NeurIPS 2022.
Response to Reviewer KcFB [3/4]
W2 / Q2: Role of ACG and training noise coverage
ACG suppresses unnecessary corrections to protect detection when poses are accurate or pose noise is small. Even with accurate input poses, feature differences across views can produce matching offsets, so applying the full PC correction can introduce errors. In the standard-training component ablation, adding ACG improves AP70 by 1.43 and 0.53 percentage points at and , respectively (Table 4). Diagnostics from the same forward pass show that ACG reduces the translation error introduced by PC from 0.1787 to 0.0247 m at zero noise (Table 5), supporting its protective role in both pose correction and detection.
Under strong pose noise, an overly conservative gate can suppress useful corrections. To examine the effect of training coverage, we kept the architecture unchanged and expanded the training noise levels from to . Both settings use 20 epochs of full-model robust fine-tuning from the same clean checkpoint. Broader training keeps 50% clean scenes and divides the remaining probability equally among the three nonzero noise levels.
Table 4. OPV2V BEV AP70 (%) and mean gate. “Without ACG” is the separately trained standard-noise ablation from the paper's Table 2.
| Training / model | AP70 ↑ | Mean gate | ||||||
|---|---|---|---|---|---|---|---|---|
| Noise | 0 | 0.2 | 0.4 | 0.6 | 0 | 0.2 | 0.4 | 0.6 |
| Standard / learned gate | 92.41 | 91.25 | 89.76 | 86.98 | 0.137 | 0.534 | 0.642 | 0.612 |
| Broad / learned gate | 92.16 | 91.09 | 90.51 | 88.40 | 0.112 | 0.578 | 0.782 | 0.808 |
| Standard / without ACG | 90.98 | 90.72 | 90.09 | 88.12 | — | — | — | — |
At , broader training raises the mean gate from 0.612 to 0.808. Raw PC translation error barely changes, while the final error falls from 0.5451 to 0.4065 m (Table 5). These results indicate that broader noise coverage helps reduce excessive suppression of useful corrections, although the final error remains higher than when applying the full PC correction.
Table 5. Mean translation error (m) ↓. PC and PC+ACG come from the same forward pass, using the raw and gated corrections, respectively. Both training settings share test-noise inputs. Bold compares PC and PC+ACG within each training/noise setting.
| Training | Input | PC | PC+ACG | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Noise | 0 | 0.2 | 0.4 | 0.6 | 0 | 0.2 | 0.4 | 0.6 | 0 | 0.2 | 0.4 | 0.6 |
| Standard | 0 | 0.3703 | 0.7426 | 1.0933 | 0.1787 | 0.1851 | 0.2303 | 0.3687 | 0.0247 | 0.1786 | 0.3023 | 0.5451 |
| Broad | 0 | 0.3703 | 0.7426 | 1.0933 | 0.1876 | 0.1820 | 0.2237 | 0.3678 | 0.0472 | 0.1809 | 0.2413 | 0.4065 |
Together, Tables 4 and 5 show that ACG protects detection by suppressing unnecessary corrections when pose noise is small. Broader noise coverage during training reduces the suppression of useful corrections at high noise, although this trade-off remains.
Response to Reviewer KcFB [1/4]
Thank you for the detailed comments on communication, methodological differences, and robustness. We address the revised questions below.
W1 / Q1 (Part 1/2): Communication-rate-matched comparisons
We adjusted the spatial feature retention ratio in Where2comm, sampling and clustering ratios in ERMVP, and mask retention ratio in InfoCom to approximately match SAComm's communication rate, with model weights fixed. Reducing communication lowers zero-noise AP70 for all three baselines on all datasets, while often improving detection under strong noise. This pattern is consistent with greater reliance on ego features: fewer sender features can weaken the benefit of collaboration at zero noise while reducing interference from misaligned observations at high noise. A flatter noise curve can therefore reflect weaker collaboration rather than better use of noisy messages. At approximately matched communication rates, SAComm retains high zero-noise accuracy and achieves the highest AP70 across the tested settings. Specifically:
Table 1. BEV AP70 (%) ↑ at original and approximately matched communication rates. Rates are Mbps per sender at 10 Hz. O and M denote Original and Matched settings.