English

Efficient Multi-View 3D Object Detection by Dynamic Token Selection and Fine-Tuning

Computer Vision and Pattern Recognition 2026-04-16 v1

Abstract

Existing multi-view three-dimensional (3D) object detection approaches widely adopt large-scale pre-trained vision transformer (ViT)-based foundation models as backbones, being computationally complex. To address this problem, current state-of-the-art (SOTA) \texttt{ToC3D} for efficient multi-view ViT-based 3D object detection employs ego-motion-based relevant token selection. However, there are two key limitations: (1) The fixed layer-individual token selection ratios limit computational efficiency during both training and inference. (2) Full end-to-end retraining of the ViT backbone is required for the multi-view 3D object detection method. In this work, we propose an image token compensator combined with a token selection for ViT backbones to accelerate multi-view 3D object detection. Unlike \texttt{ToC3D}, our approach enables dynamic layer-wise token selection within the ViT backbone. Furthermore, we introduce a parameter-efficient fine-tuning strategy, which trains only the proposed modules, thereby reducing the number of fine-tuned parameters from more than 300300 million (M) to only 1.61.6 M. Experiments on the large-scale NuScenes dataset across three multi-view 3D object detection approaches demonstrate that our proposed method decreases computational complexity (GFLOPs) by 48%48\% ... 55%55\%, inference latency (on an \texttt{NVIDIA-GV100} GPU) by 9%9\% ... 25%25\%, while still improving mean average precision by 1.0%1.0\% ... 2.8%2.8\% absolute and NuScenes detection score by 0.4%0.4\% ... 1.2%1.2\% absolute compared to so-far SOTA \texttt{ToC3D}.

Keywords

Cite

@article{arxiv.2604.13586,
  title  = {Efficient Multi-View 3D Object Detection by Dynamic Token Selection and Fine-Tuning},
  author = {Danish Nazir and Antoine Hanna-Asaad and Lucas Görnhardt and Jan Piewek and Thorsten Bagdonat and Tim Fingscheidt},
  journal= {arXiv preprint arXiv:2604.13586},
  year   = {2026}
}
R2 v1 2026-07-01T12:10:17.673Z