# C$^2$-Routing: Complementary Cropping and Compression for Efficient Vision-Language Inference

Source: [arXiv](https://arxiv.org/abs/2611.14469v1)  
Feed7 permalink: https://feed7.dev/p/2611-14469v1-1um73ov  
Published: 2026-11-10T18:32:28.000Z  
Trust: Needs Review (needs_review)

## Why Included

C²-Routing adaptively crops or compresses images before vision-language inference, with a fallback gate for risky reductions. Its reported savings are measured against already token-capped inputs across three models and 18 datasets.

## Source Summary

C²-Routing chooses between cropping for local detail and compression for broader context, with a safety gate that can fall back to the original input. Across **three VLMs** and **18 datasets**, it reduced FLOPs by 36.53% and visual tokens by 42.36% on average, with a 0.77% mean relative task-score decrease.

## Practical Implication

For agents that inspect screenshots, documents, or other images, adaptive preprocessing could reduce inference work while preserving relevant detail. The safety gate is part of the approach, so a simple fixed token cap may not capture the same tradeoff.

## Agent-Ready Context

C²-Routing chooses between cropping for local detail and compression for broader context, with a safety gate that can fall back to the original input. Across **three VLMs** and **18 datasets**, it reduced FLOPs by 36.53% and visual tokens by 42.36% on average, with a 0.77% mean relative task-score decrease.

For agents that inspect screenshots, documents, or other images, adaptive preprocessing could reduce inference work while preserving relevant detail. The safety gate is part of the approach, so a simple fixed token cap may not capture the same tradeoff.

The comparisons used a full-image baseline capped at 1,024 visual tokens, and experiments ran on RTX 3090 GPUs. The material does not establish how the method performs on other workloads or hardware.

## Connected Context

Feed7 judgment across 1059 accumulated Signals:

The results make adaptive image preprocessing a measurable option for reducing visual inference work, with the reported average task-score decrease and a fallback gate defining the tradeoff. They support testing the approach on screenshot or document workflows, but do not establish transfer beyond the studied models, datasets, token cap, and hardware.

- [ReToken: One Token to Improve Vision-Language Models for Visual Retrieval](https://feed7.dev/p/2607-28627v1-0pc8021) — C²-Routing reduces input cost by choosing cropping or compression before inference; ReToken selects relevant visual tokens from cached context. Both limit visual evidence processed, but at different stages and with different mechanisms.

## Context Map

- Layer: context
- Domains: image
- Topics: context-engineering, model-selection

## Uncertainty

- The comparisons used a full-image baseline capped at 1,024 visual tokens, and experiments ran on RTX 3090 GPUs. The material does not establish how the method performs on other workloads or hardware.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.
