用张量余弦积压缩注意力机制，显著降低视觉Transformer的计算开销。
A Computationally Efficient Multidimensional Vision Transformer
- 基于张量余弦积构建新架构，利用图像多线性结构提升效率。
- 参数量减少1/C（C为通道数），分类与分割任务精度仍保持领先。
- 适合需要轻量化部署的视觉模型开发者参考。
视觉Transformer在众多计算机视觉任务中表现卓越，但其实际部署受限于高计算和内存开销。本文提出一种基于张量余弦积（Cproduct）的新型张量框架，通过挖掘图像数据的多线性结构及余弦变换的正交性，实现高效的注意力机制与结构化特征表示。我们建立了张量余弦积的理论基础，分析其代数性质，并构建了基于Cproduct的视觉Transformer架构（TCP-ViT）。在标准分类与分割基准上的数值实验表明，该方法在保持竞争力精度的同时，实现了统一的1/C参数缩减（C为通道数）。
原文摘要 · Abstract (English)
Vision Transformers have achieved state-of-the-art performance in a wide range of computer vision tasks, but their practical deployment is limited by high computational and memory costs. In this paper, we introduce a novel tensor-based framework for Vision Transformers built upon the Tensor Cosine Product (Cproduct). By exploiting multilinear structures inherent in image data and the orthogonality of cosine transforms, the proposed approach enables efficient attention mechanisms and structured feature representations. We develop the theoretical foundations of the tensor cosine product, analyze its algebraic properties, and integrate it into a new Cproduct-based Vision Transformer architecture (TCP-ViT). Numerical experiments on standard classification and segmentation benchmarks demonstrate that the proposed method achieves a uniform 1/C parameter reduction (where C is the number of channels) while maintaining competitive accuracy.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。