[ECCV 2026] SFPruner: Structured Redundancy Modeling for Efficient Visual Token Pruning in High-Resolution MLLMs

Juwon Song¹, Woohyeong Kim², Kyeongbo Kong^2
1LG Electronics    ²Pusan National University

Project page  |  arXiv: coming soon  |  Code: coming soon

SFPruner Algorithm

SFPruner consists of three stages:

1. Semantic guidance
   Estimate base token importance using instruction relevance and visual saliency.

2. Semantics-Guided Ridge Leverage Score (SG-RLS)
   Suppress globally redundant covariance directions by reweighting token scores with ridge leverage-based structural information.

3. Ranking-Based Directional Masking
   Resolve residual pairwise redundancy by allowing higher-scoring tokens to suppress similar lower-scoring alternatives through a single parallel masking operation.

This design avoids the iterative dependency of subset-optimization methods such as DPP-, diversity-, or graph-based selection, enabling direct Top-K pruning with stable selection latency.