Researchers have developed SAMIC, a novel method for efficient perceptual image compression that utilizes Mamba, a state space model known for its long-range modeling capabilities and linear complexity. Unlike traditional methods that can struggle with semantic continuity, SAMIC introduces a semantic-aware Mamba block (SAMB) that allows scanning to be guided by semantic features. Additionally, an SVD-inspired redundancy reduction module (SVD-RRM) is incorporated to reduce channel-wise redundancy in latent features. Experiments indicate that SAMIC outperforms current state-of-the-art approaches in balancing rate, distortion, and perception while maintaining lower model complexity. AI
Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →
IMPACT Introduces a more efficient approach to image compression by leveraging Mamba, potentially improving quality at lower bitrates.
RANK_REASON This is a research paper detailing a new method for image compression.