SVL: Spike-based Vision-language Pretraining for Efficient 3D Open-world Understanding

TL;DR

The paper introduces SVL, a spike-based vision-language pretraining framework that enhances the 3D understanding capabilities of SNNs, achieving state-of-the-art results in zero-shot classification and multimodal tasks.

Contribution

SVL is the first scalable, generalizable, and hardware-friendly pretraining framework that significantly improves SNNs' performance in open-world 3D understanding tasks.

Findings

Achieves 85.4% top-1 accuracy in zero-shot 3D classification.

Outperforms prior SNNs on downstream tasks like 3D classification, DVS action recognition, detection, and segmentation.

Enables SNNs to perform open-world 3D question answering, sometimes surpassing ANNs.

Abstract

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.