Distributed Compression Coding Based on Convolutional Sparse Coding Using Multiple Key Frames
Yosuke Higuchi, Muhammad Ajmal Muhaimin Bin Mustafa, Yoshimitsu Kuroki · 2024
This paper proposes an approach to video compression using Convolutional Sparse Coding (CSC) with a focus on differentiating between Key Frames (KF) and Non-Key Frames (NKF). We introduce a Consensus Framework for CSC and apply the Alternating Direction Method of Multipliers (ADMM) for optimization. Our method shows improved compression efficiency and image quality, particularly for NKFs. Experimental results on the Foreman dataset demonstrate enhanced PSNR and SSIM metrics compared to traditional methods, especially at lower measurement rates.