# Revisiting Token Compression for Accelerating ViT-based Sparse Multi-View 3D Object Detectors

arXiv: https://arxiv.org/abs/2604.14563 · HTML: https://mlsanity.com/paper/2604.14563/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Token pruning and merging for ViTs, one of 12 directions in Efficient LLM inference and serving.
- Builds on 48 of its 57 references that are in the map.
- First posted on arXiv 16 Apr 2026.

## Placed in

- Efficient LLM inference and serving / [Token pruning and merging for ViTs](https://mlsanity.com/inference/token-pruning-and-merging-for-vits/)

## Builds on (most cited first, 30 of 48)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [nuScenes: A multimodal dataset for autonomous driving](https://mlsanity.com/paper/1903.11027/) (2019)
- [Categorical Reparameterization with Gumbel-Softmax](https://mlsanity.com/paper/1611.01144/) (2016)
- [BEVFormer: Learning Bird's-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers](https://mlsanity.com/paper/2203.17270/) (2022)
- [Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet](https://mlsanity.com/paper/2101.11986/) (2021)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [CvT: Introducing Convolutions to Vision Transformers](https://mlsanity.com/paper/2103.15808/) (2021)
- [DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification](https://mlsanity.com/paper/2106.02034/) (2021)
- [Argoverse 2: Next Generation Datasets for Self-Driving Perception and Forecasting](https://mlsanity.com/paper/2301.00493/) (2023)
- [DETR3D: 3D Object Detection from Multi-view Images via 3D-to-2D Queries](https://mlsanity.com/paper/2110.06922/) (2021)
- [BEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object Detection](https://mlsanity.com/paper/2206.10092/) (2022)
- [PETR: Position Embedding Transformation for Multi-View 3D Object Detection](https://mlsanity.com/paper/2203.05625/) (2022)
- [EVA-02: A Visual Representation for Neon Genesis](https://mlsanity.com/paper/2303.11331/) (2023)
- [Not All Patches are What You Need: Expediting Vision Transformers via Token Reorganizations](https://mlsanity.com/paper/2202.07800/) (2022)
- [BEVDet4D: Exploit Temporal Cues in Multi-camera 3D Object Detection](https://mlsanity.com/paper/2203.17054/) (2022)
- [Exploring Object-Centric Temporal Modeling for Efficient Multi-View 3D Object Detection](https://mlsanity.com/paper/2303.11926/) (2023)
- [Time Will Tell: New Outlooks and A Baseline for Temporal Multi-View 3D Object Detection](https://mlsanity.com/paper/2210.02443/) (2022)
- [Evo-ViT: Slow-Fast Token Evolution for Dynamic Vision Transformer](https://mlsanity.com/paper/2108.01390/) (2021)
- [An Energy and GPU-Computation Efficient Backbone Network for Real-Time Object Detection](https://mlsanity.com/paper/1904.09730/) (2019)
- [Sparse4D: Multi-view 3D Object Detection with Sparse Spatial-Temporal Fusion](https://mlsanity.com/paper/2211.10581/) (2022)
- [FlexiViT: One Model for All Patch Sizes](https://mlsanity.com/paper/2212.08013/) (2022)
- [SparseBEV: High-Performance Sparse 3D Object Detection from Multi-Camera Videos](https://mlsanity.com/paper/2308.09244/) (2023)
- [Token Fusion: Bridging the Gap between Token Pruning and Token Merging](https://mlsanity.com/paper/2312.01026/) (2023)
- [Not All Tokens Are Equal: Human-centric Visual Analysis via Token Clustering Transformer](https://mlsanity.com/paper/2204.08680/) (2022)
- [Sparse DETR: Efficient End-to-End Object Detection with Learnable Sparsity](https://mlsanity.com/paper/2111.14330/) (2021)
- [World4Drive: End-to-End Autonomous Driving via Intention-aware Physical Latent World Model](https://mlsanity.com/paper/2507.00603/) (2025)
- [Sparse4D v2: Recurrent Temporal Fusion with Sparse Model](https://mlsanity.com/paper/2305.14018/) (2023)

## Built on by (most cited first, 1 of 1)

- [Variable-Granularity Tokenization for High-Resolution Object Detection](https://mlsanity.com/paper/2608.28706/) (2026)

## Abstract (arXiv)

Vision Transformer (ViT)-based sparse multi-view 3D object detectors have achieved remarkable accuracy but still suffer from high inference latency due to heavy token processing. To accelerate these models, token compression has been widely explored. However, our revisit of existing strategies, such as token pruning, merging, and patch size enlargement, reveals that they often discard informative background cues, disrupt contextual consistency, and lose fine-grained semantics, negatively affecting 3D detection. To overcome these limitations, we propose SEPatch3D, a novel framework that dynamically adjusts patch sizes while preserving critical semantic information within coarse patches. Specifically, we design Spatiotemporal-aware Patch Size Selection (SPSS) that assigns small patches to scenes containing nearby objects to preserve fine details and large patches to background-dominated scenes to reduce computation cost. To further mitigate potential detail loss, Informative Patch Selection (IPS) selects the informative patches for feature refinement, and Cross-Granularity Feature Enhancement (CGFE) injects fine-grained details into selected coarse patches, enriching semantic features. Experiments on the nuScenes and Argoverse 2 validation sets show that SEPatch3D achieves up to 57% faster inference than the StreamPETR baseline and 20% higher efficiency than the state-of-the-art ToC3D-faster, while preserving comparable detection accuracy. Code is available at https://github.com/Mingqj/SEPatch3D.