3D-BBDA is a backpropagation-free black-box domain adaptation framework for 3D point cloud classification that requires neither source data nor source model internals. The method combines pseudo-label-guided prototype alignment with a Gaussian-perturbed prototype prior completion strategy to stabilize adaptation under sparse pseudo-labels, and a weighted prediction module to balance source priors with target-specific correction. Experiments on multiple corrupted 3D benchmarks demonstrate consistent robustness gains with minimal overhead, enabling real-time and privacy-preserving deployment. Robustness and real-time feasibility are validated via standard corruption benchmarks and inference overhead measurements.
Detailed Comparison with BFTT3D
BFTT3D is the closest prior method to 3D-BBDA because both avoid test-time backpropagation and employ non-parametric target representations together with prototype-based shared-subspace alignment. The primary methodological distinction is therefore not the introduction of a completely new subspace-alignment objective, but the reformulation of this pipeline under a stricter source-free, logits-only black-box interface.
3D-BBDA does not claim the shared-subspace alignment objective itself as a new contribution. Instead, the method changes the information available to this alignment: source-derived memory is removed and replaced by pseudo-label-guided target prototypes, missing classes are completed from an unlabeled Gaussian-perturbed target prior, and the resulting alignment prediction is fused with the black-box source prediction. These changes enable the alignment framework to operate under the joint source-free, black-box, and backpropagation-free constraints.
Algorithm
The complete 3D-BBDA inference procedure first constructs a fixed target-derived prototype memory from accumulated features and pseudo-labels, then performs block-wise shared-subspace alignment and weighted prediction fusion.
Details of Controlled Recent Baseline Ports
Recent adaptation methods differ substantially in the information they assume available at test time. We therefore distinguish native implementations from controlled ports and explicitly document the information-access and architecture assumptions used for each comparison.
Controlled PURGE-GATE Port
Native PURGE-GATE PG-SF is designed for Transformer point-cloud architectures and relies on CLS and attention Q/K representations. These quantities are not available in PointNet or GDANet and, more importantly, would not be accessible through the logits-only black-box interface assumed by 3D-BBDA.
For the PointNet/GDANet comparison, we therefore construct a controlled PG-SP port. Purging is applied to the earliest 64-dimensional point-wise representation available in each backbone, and source-derived embedding statistics are used according to the PG-SP formulation. No target-side backpropagation is performed.
Implementation Details
Native PURGE-GATE PG-SF operates on Transformer-specific CLS/Q/K representations and is therefore not directly applicable to PointNet/GDANet. We implement a controlled PG-SP port by applying purging to the earliest 64-dimensional point-wise representation using source-derived embedding statistics.
DINE-3D denotes a controlled feature-space port of DINE, originally proposed for 2D black-box domain adaptation. It uses the same frozen PointNN target features as 3D-BBDA while retaining DINE’s source-probability distillation, EMA memory, information-maximization, MixUp, and target fine-tuning objectives. Unlike 3D-BBDA, DINE-3D requires target-side backpropagation.
Conclusion
We proposed 3D-BBDA, a backpropagation-free black-box domain adaptation framework for 3D point cloud classification that requires neither source data nor source model internals. The method combines pseudo-label-guided prototype alignment with a Gaussian-perturbed prototype prior completion strategy to stabilize adaptation under sparse pseudo-labels, and a weighted prediction module to balance source priors with target-specific correction. Experiments on multiple corrupted 3D benchmarks demonstrate consistent robustness gains with minimal overhead, enabling real-time and privacy-preserving deployment. We validate robustness and real-time feasibility via standard corruption benchmarks and inference overhead measurements.
Acknowledgments
This work was supported by the Indian Space Research Organisation under Grant No. RespondAI, P.No.ISRO/RES/3/1147/26-27. Shivangi Rai gratefully acknowledges the financial support provided through the Visvesvaraya PhD Fellowship by the Ministry of Electronics and Information Technology (MeitY), Government of India.
Frequently Asked Questions
What problem does 3D-BBDA address? It addresses domain adaptation for 3D point cloud classification under source-free, black-box, and backpropagation-free constraints.
Does 3D-BBDA require access to source data or source model internals? No. The framework requires neither source data nor source model internals.
How does 3D-BBDA handle sparse pseudo-labels? It uses Gaussian-perturbed prototype prior completion to stabilize adaptation when pseudo-labels are sparse.
How does DINE-3D differ from 3D-BBDA? DINE-3D retains DINE’s target fine-tuning objectives and requires target-side backpropagation, unlike 3D-BBDA.