New Score-Based Turbo Message Passing Algorithm Enhances Compressive Imaging

Researchers have developed a new message-passing algorithm called Score-Based Turbo Message Passing (STMP) for compressive imaging. This method integrates score-based generative models with empirical Bayes denoising to improve image reconstruction, particularly in underdetermined scenarios. For systems with quantized measurements, a variant called Quantized STMP (Q-STMP) was also introduced, which includes a dequantization module and remains robust even with 1-bit quantization. Experiments show that STMP offers a better performance-complexity tradeoff than existing methods and typically converges within 10 iterations. AI

IMPACT This new algorithm could improve image reconstruction quality and efficiency in various imaging applications.

RANK_REASON The cluster contains a research paper detailing a new algorithm for compressive imaging. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.CV →

• Chang Cai
• Compressive imaging and dual moire laser interferometer as metrology tools
• FFHQ dataset
• mini–mental state examination
• Q-STMP
• Quantized STMP
• Score-Based Turbo Message Passing
• STEAP2

• paper
• infra

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