New ConFusion framework enables fine-grained control over image fusion

Researchers have developed ConFusion, a new framework for controllable infrared and visible image fusion. This method addresses limitations in existing approaches by learning a continuous fusion space, allowing for fine-grained, instance-level modulation of fused images. ConFusion utilizes a dual-branch architecture and Gaussian-conditioned spatial-aware modulation to disentangle representations and enhance semantic consistency. The framework can parse user intents from multimodal large language models to guide the fusion process, achieving state-of-the-art performance in fusion quality and downstream tasks. AI

IMPACT This research could lead to more adaptable and precise image processing tools for various applications.

RANK_REASON The cluster contains an academic paper detailing a new method for image fusion. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.CV →

• ConFusion
• Gaussian-conditioned spatial-aware modulation
• Grounded SAM-based instance masks
• Mask-Guided Specific Feature Modulator
• multimodal large language model
• Text-Driven Invariant Feature Enhancer

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →