# CARVE: Cross-Slice Anisotropic Reallocation of Visual Evidence for Efficient 3D Medical Volume Understanding

arXiv: https://arxiv.org/abs/2608.04515 · HTML: https://mlsanity.com/paper/2608.04515/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Visual token pruning in multimodal models, one of 12 directions in Efficient LLM inference and serving.
- Builds on 20 of its 37 references that are in the map.
- First posted on arXiv 5 Aug 2026.

## Placed in

- Efficient LLM inference and serving / [Visual token pruning in multimodal models](https://mlsanity.com/inference/visual-token-pruning-in-multimodal-models/)

## Builds on (most cited first, 20 of 20)

- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Day](https://mlsanity.com/paper/2306.00890/) (2023)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [Not All Patches are What You Need: Expediting Vision Transformers via Token Reorganizations](https://mlsanity.com/paper/2202.07800/) (2022)
- [Generalist Foundation Models from a Multimodal Dataset for 3D Computed Tomography](https://mlsanity.com/paper/2403.17834/) (2024)
- [Hulu-Med: A Transparent Generalist Model towards Holistic Medical Vision-Language Understanding](https://mlsanity.com/paper/2510.08668/) (2025)
- [CT2Rep: Automated Radiology Report Generation for 3D Medical Imaging](https://mlsanity.com/paper/2403.06801/) (2024)
- [Beyond Attention or Similarity: Maximizing Conditional Diversity for Token Pruning in MLLMs](https://mlsanity.com/paper/2506.10967/) (2025)
- [PACT: Pruning and Clustering-Based Token Reduction for Faster Visual Language Models](https://mlsanity.com/paper/2504.08966/) (2025)
- [What Kind of Visual Tokens Do We Need? Training-free Visual Token Pruning for Multi-modal Large Language Models from the Perspective of Graph](https://mlsanity.com/paper/2501.02268/) (2025)
- [MedFrameQA: A Multi-Image Medical VQA Benchmark for Clinical Reasoning](https://mlsanity.com/paper/2505.16964/) (2025)
- [MMTok: Multimodal Coverage Maximization for Efficient Inference of VLMs](https://mlsanity.com/paper/2508.18264/) (2025)
- [Multi-Granular Spatio-Temporal Token Merging for Training-Free Acceleration of Video LLMs](https://mlsanity.com/paper/2507.07990/) (2025)
- [3D-RAD: A Comprehensive 3D Radiology Med-VQA Dataset with Multi-Temporal Analysis and Diverse Diagnostic Tasks](https://mlsanity.com/paper/2506.11147/) (2025)
- [Read Like a Radiologist: Efficient Vision-Language Model for 3D Medical Imaging Interpretation](https://mlsanity.com/paper/2412.13558/) (2024)
- [OmniV-Med: Scaling Medical Vision-Language Model for Universal Visual Understanding](https://mlsanity.com/paper/2504.14692/) (2025)
- [MMG-Vid: Maximizing Marginal Gains at Segment-level and Token-level for Efficient Video LLMs](https://mlsanity.com/paper/2508.21044/) (2025)
- [Geometry-Guided 3D Visual Token Pruning for Video-Language Models](https://mlsanity.com/paper/2604.18260/) (2026)
- [E-MRL: Cross-view Aligned Evidence-driven Multimodal Reinforcement Learning for Reliable 3D Tumor Analysis](https://mlsanity.com/paper/2606.23888/) (2026)
- [Brain-Adapter: A Dual-Stream Vision-Language MIL Framework for Comprehensive 3D CT Diagnosis of Acute Intracranial Pathologies](https://mlsanity.com/paper/2606.23494/) (2026)

## Abstract (arXiv)

Slice-based MLLMs leverage mature 2D encoders by representing 3D volumes as sequences of 2D slices. However, this slice-wise formulation produces thousands of visual tokens that burden the LLM backbone, many of which capture overlapping visual evidence across adjacent slices. To understand how effectively a growing visual token budget improves performance, we perform scaling analyses on two 3D medical VQA benchmarks and find diminishing returns: cost keeps rising while accuracy saturates, and improving in-plane resolution is more effective than adding slices at comparable budgets. The budget should therefore be allocated more selectively rather than simply enlarged, yet most token compression methods are designed for 2D images or videos, where redundancy arises from spatial layout or temporal motion rather than from near-duplicate content along the depth axis. We present CARVE, a training-free framework that compresses visual tokens prior to LLM inference and casts token reduction as budget-constrained 2.5D allocation. CARVE partitions the depth axis into coherent windows and allocates tokens non-uniformly according to normalized cross-slice evidence. Under a shared budget, CARVE builds spatial anchors on representative slices and retrieves locally varying evidence from the full volume, then merges remaining eligible tokens into nearby anchors within each window. Removing roughly 80% of the visual tokens on Hulu-Med-7B, CARVE leads all compression baselines on every AMOS-MM report-generation metric, with 6.2 points higher retention of full-token quality than the strongest baseline, and preserves 98.1% of full-token performance across three VQA benchmarks.