Accelerate 3D Object Detection Models via Zero-Shot Attention Key Pruning

• URL: http://arxiv.org/abs/2503.08101v2
• Date: Wed, 12 Mar 2025 04:03:24 GMT
• Title: Accelerate 3D Object Detection Models via Zero-Shot Attention Key Pruning
• Authors: Lizhen Xu, Xiuxiu Bai, Xiaojun Jia, Jianwu Fang, Shanmin Pang,
• Abstract summary: We propose a zero-shot runtime pruning method for transformer decoders in 3D object detection models.<n>Our method achieves a 1.99x speedup in the transformer decoder of the latest ToC3D model, with only a minimal performance loss of less than 1%.
• Score: 15.40654753734657
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Query-based methods with dense features have demonstrated remarkable success in 3D object detection tasks. However, the computational demands of these models, particularly with large image sizes and multiple transformer layers, pose significant challenges for efficient running on edge devices. Existing pruning and distillation methods either need retraining or are designed for ViT models, which are hard to migrate to 3D detectors. To address this issue, we propose a zero-shot runtime pruning method for transformer decoders in 3D object detection models. The method, termed tgGBC (trim keys gradually Guided By Classification scores), systematically trims keys in transformer modules based on their importance. We expand the classification score to multiply it with the attention map to get the importance score of each key and then prune certain keys after each transformer layer according to their importance scores. Our method achieves a 1.99x speedup in the transformer decoder of the latest ToC3D model, with only a minimal performance loss of less than 1%. Interestingly, for certain models, our method even enhances their performance. Moreover, we deploy 3D detectors with tgGBC on an edge device, further validating the effectiveness of our method. The code can be found at https://github.com/iseri27/tg_gbc.

Related papers

• Cubify Anything: Scaling Indoor 3D Object Detection [4.338330763853994]
  We consider indoor 3D object detection with respect to a single RGB(-D) frame acquired from a commodity handheld device. We introduce the Cubify-Anything 1M dataset, which exhaustively labels over 400K 3D objects on over 1K highly accurate laser-scanned scenes. Next, we establish Cubify Transformer (CuTR), a fully Transformer 3D object detection baseline which rather than operating in 3D on point or voxel-based representations, predicts 3D boxes directly from 2D features derived from RGB(-D) inputs.
  arXiv  Detail & Related papers  (2024-12-05T18:59:09Z)
• 3DGS-CD: 3D Gaussian Splatting-based Change Detection for Physical Object Rearrangement [2.2122801766964795]
  We present 3DGS-CD, the first 3D Gaussian Splatting (3DGS)-based method for detecting physical object rearrangements in 3D scenes. Our approach estimates 3D object-level changes by comparing two sets of unaligned images taken at different times. Our method can accurately identify changes in cluttered environments using sparse (as few as one) post-change images within as little as 18s.
  arXiv  Detail & Related papers  (2024-11-06T07:08:41Z)
• Make Your ViT-based Multi-view 3D Detectors Faster via Token Compression [78.93023152602525]
  Slow inference speed is one of the most crucial concerns for deploying multi-view 3D detectors to tasks with high real-time requirements like autonomous driving. We propose a simple yet effective method called TokenCompression3D (ToC3D) Our method can nearly maintain the performance of recent SOTA with up to 30% inference speedup, and the improvements are consistent after scaling up the ViT and input resolution.
  arXiv  Detail & Related papers  (2024-09-01T06:58:08Z)
• WidthFormer: Toward Efficient Transformer-based BEV View Transformation [21.10523575080856]
  WidthFormer is a transformer-based module to compute Bird's-Eye-View (BEV) representations from multi-view cameras for real-time autonomous-driving applications. We first introduce a novel 3D positional encoding mechanism capable of accurately encapsulating 3D geometric information. We then develop two modules to compensate for potential information loss due to feature compression.
  arXiv  Detail & Related papers  (2024-01-08T11:50:23Z)
• Hourglass Tokenizer for Efficient Transformer-Based 3D Human Pose Estimation [73.31524865643709]
  We present a plug-and-play pruning-and-recovering framework, called Hourglass Tokenizer (HoT), for efficient transformer-based 3D pose estimation from videos. Our HoDT begins with pruning pose tokens of redundant frames and ends with recovering full-length tokens, resulting in a few pose tokens in the intermediate transformer blocks. Our method can achieve both high efficiency and estimation accuracy compared to the original VPT models.
  arXiv  Detail & Related papers  (2023-11-20T18:59:51Z)
• S$^3$-MonoDETR: Supervised Shape&Scale-perceptive Deformable Transformer for Monocular 3D Object Detection [21.96072831561483]
  This paper proposes a novel Supervised Shape&Scale-perceptive Deformable Attention'' (S$3$-DA) module for monocular 3D object detection. Benefiting from this, S$3$-DA effectively estimates receptive fields for query points belonging to any category, enabling them to generate robust query features. Experiments on KITTI and Open datasets demonstrate that S$3$-DA significantly improves the detection accuracy.
  arXiv  Detail & Related papers  (2023-09-02T12:36:38Z)
• 3D Small Object Detection with Dynamic Spatial Pruning [62.72638845817799]
  We propose an efficient feature pruning strategy for 3D small object detection. We present a multi-level 3D detector named DSPDet3D which benefits from high spatial resolution. It takes less than 2s to directly process a whole building consisting of more than 4500k points while detecting out almost all objects.
  arXiv  Detail & Related papers  (2023-05-05T17:57:04Z)
• Hierarchical Point Attention for Indoor 3D Object Detection [111.04397308495618]
  This work proposes two novel attention operations as generic hierarchical designs for point-based transformer detectors. First, we propose Multi-Scale Attention (MS-A) that builds multi-scale tokens from a single-scale input feature to enable more fine-grained feature learning. Second, we propose Size-Adaptive Local Attention (Local-A) with adaptive attention regions for localized feature aggregation within bounding box proposals.
  arXiv  Detail & Related papers  (2023-01-06T18:52:12Z)
• Embracing Single Stride 3D Object Detector with Sparse Transformer [63.179720817019096]
  In LiDAR-based 3D object detection for autonomous driving, the ratio of the object size to input scene size is significantly smaller compared to 2D detection cases. Many 3D detectors directly follow the common practice of 2D detectors, which downsample the feature maps even after quantizing the point clouds. We propose Single-stride Sparse Transformer (SST) to maintain the original resolution from the beginning to the end of the network.
  arXiv  Detail & Related papers  (2021-12-13T02:12:02Z)
• Improving 3D Object Detection with Channel-wise Transformer [58.668922561622466]
  We propose a two-stage 3D object detection framework (CT3D) with minimal hand-crafted design. CT3D simultaneously performs proposal-aware embedding and channel-wise context aggregation. It achieves the AP of 81.77% in the moderate car category on the KITTI test 3D detection benchmark.
  arXiv  Detail & Related papers  (2021-08-23T02:03:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.