Dimensions overall score 4.0

GitHub Code Pulse

No public code linked for this paper yet.

• SAFT integrates a Point-BERT-inspired encoder, a sensitivity-guided token filtering (STF) unit, a quantization block, and an SNR-aware decoder.method 95%
• The sensitivity-guided token filtering (STF) module assigns token-wise importance scores based on the reconstruction sensitivity of each token under channel perturbation.method 95%
• SAFT improves geometric fidelity (D1/D2 PSNR) compared with a separate source-channel coding pipeline (G-PCC combined with LDPC and QAM) and existing learned baselines.result 90%
• The largest gains in geometric fidelity are observed in low-SNR regimes, highlighting improved robustness under limited bandwidth.result 90%
• A training-only symbol-usage penalty is employed to stabilize the discrete representation without affecting the transmitted payload.technical 90%
• SAFT utilizes a Point-BERT-inspired transformer for encoding the input point cloud into patch-level latent tokens.method 95%
• The STF module uses a lightweight MLP to assign token-wise sensitivity scores.method 90%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card