MixPT is a point cloud transformer that interleaves window and global attention blocks while keeping a fixed window serialization pattern. It applies global attention to pooled tokens for large scenes, feeding residual updates back to the full-resolution stream. Experiments on indoor and outdoor benchmarks show MixPT matches or exceeds the accuracy of existing window or global attention models while achieving lower latency, and it can be scaled by repeating the interleaved pattern.
arXiv:2609.23733v1 Announce Type: new
Abstract: Feed-forward visual geometry models such as the Visual Geometry Grounded Transformer (VGGT) have recently enabled direct 3D reconstruction from multi-v...
By Abteen Arab, Guile Wu, Chengjie Huang, Dongfeng Bai
The paper introduces Hardware‑aware Calibrated Clustered Attention (BC attention) to speed up the global attention layers of the Visual Geometry Grounded Transformer (VGGT) used for 3D scene reconstruction. By clustering queries within hardware‑friendly neighborhood blocks and adding hashing hyperplane calibration plus threshold‑based error compensation, BC attention reduces computation and memory traffic, enabling it to handle long sequences with practical latency gains on GPUs. Experiments show that calibrated BC attention accelerates global attention by 2.10–2.63× and the entire backbone by 1.77–2.35×, with only a 1% loss in accuracy, and achieves up to 2.87× latency improvement with less than 5% performance drop.
By Weitian Wang, Shubham Rai, Cecilia De La Parra, Akash Kumar
LitePT introduces a lighter yet stronger 3D point cloud backbone that strategically uses convolutional layers in early, high‑resolution stages and switches to attention in deeper, low‑resolution layers. By adding a novel, parameter‑free 3D positional encoding called PointROPE, LitePT preserves spatial layout while reducing redundancy. The resulting model has 3.6× fewer parameters, runs twice as fast, and uses half the memory of the state‑of‑the‑art Point Transformer V3, yet matches or surpasses its performance across multiple tasks and datasets.
By Yuanwen Yue, Damien Robert, Jianyuan Wang, Sunghwan Hong, Jan Dirk Wegner, Christian Rupprecht, Konrad Schindler
arXiv:2605.12491v2 Announce Type: replace
Abstract: Vision Transformers (ViTs) learn rich visual-semantic representations through all-to-all self-attention among patch tokens. However, this design im...
By Alan Z. Song, Yinjie Chen, Mu Nan, Deva Ramanan, Michael J. Tarr, Andrew F. Luo
BlazeVGGT is a training‑free, plug‑and‑play module that speeds up the VGGT visual geometry model by sparsifying cross‑view attention. It selects a compact set of complementary views using a greedy determinantal point process, preserving dense per‑frame predictions while eliminating redundant computations. On 2,000‑frame sequences, BlazeVGGT achieves roughly 13× faster inference and improves Chamfer Distance from 0.41 to 0.32, demonstrating a strong accuracy–efficiency trade‑off for point‑cloud reconstruction and camera‑pose estimation.
By Shujun Yao, Junhui Lin, Congying Han, Tiande Guo, Ziwen Liu, Bonan Li