SVL: Spike-based Vision-language Pretraining for Efficient 3D Open-world Understanding
About
Spiking Neural Networks (SNNs) provide an energy-efficient way to extract 3D spatio-temporal features. However, existing SNNs still exhibit a significant performance gap compared to Artificial Neural Networks (ANNs) due to inadequate pre-training strategies. These limitations manifest as restricted generalization ability, task specificity, and a lack of multimodal understanding, particularly in challenging tasks such as multimodal question answering and zero-shot 3D classification. To overcome these challenges, we propose a Spike-based Vision-Language (SVL) pretraining framework that empowers SNNs with open-world 3D understanding while maintaining spike-driven efficiency. SVL introduces two key components: (i) Multi-scale Triple Alignment (MTA) for label-free triplet-based contrastive learning across 3D, image, and text modalities, and (ii) Re-parameterizable Vision-Language Integration (Rep-VLI) to enable lightweight inference without relying on large text encoders. Extensive experiments show that SVL achieves a top-1 accuracy of 85.4% in zero-shot 3D classification, surpassing advanced ANN models, and consistently outperforms prior SNNs on downstream tasks, including 3D classification (+6.1%), DVS action recognition (+2.1%), 3D detection (+1.1%), and 3D segmentation (+2.1%) with remarkable efficiency. Moreover, SVL enables SNNs to perform open-world 3D question answering, sometimes outperforming ANNs. To the best of our knowledge, SVL represents the first scalable, generalizable, and hardware-friendly paradigm for 3D open-world understanding, effectively bridging the gap between SNNs and ANNs in complex open-world understanding tasks. Code is available https://github.com/bollossom/SVL.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| 3D Object Classification | ModelNet40 (test) | Accuracy92.6 | 321 | |
| 3D Object Classification | ModelNet40 | -- | 89 | |
| 3D Semantic Segmentation | SemanticKITTI (val) | mIoU69.7 | 75 | |
| 3D Object Classification | ScanObjectNN | Accuracy83.4 | 35 | |
| 4D semantic segmentation | Synthia 4D | mIoU80.05 | 31 | |
| Action Recognition | DVS128Gesture | Accuracy98.5 | 18 | |
| Zero-shot Classification | ModelNet40 | Accuracy83.1 | 18 | |
| 3D Object Detection | KITTI | AP (Easy)90.7 | 17 | |
| 3D Zero-shot classification | Objaverse-LVIS zero-shot | Accuracy43.4 | 15 | |
| 3D Classification | ScanObjectNN Scan-O (test) | Accuracy82.1 | 12 |