Deep Inertia $L_p$ Half-Quadratic Splitting Unrolling Network for Sparse View CT Reconstruction
This addresses a challenging inverse problem in medical imaging, offering incremental improvements for CT reconstruction under sparse data conditions.
The paper tackles sparse view CT reconstruction by developing an inertial Lp-norm half-quadratic splitting algorithm with deep unrolling, proving convergence and showing it outperforms existing methods in scenarios with fewer views and complex noise.
Sparse view computed tomography (CT) reconstruction poses a challenging ill-posed inverse problem, necessitating effective regularization techniques. In this letter, we employ $L_p$-norm ($0<p<1$) regularization to induce sparsity and introduce inertial steps, leading to the development of the inertial $L_p$-norm half-quadratic splitting algorithm. We rigorously prove the convergence of this algorithm. Furthermore, we leverage deep learning to initialize the conjugate gradient method, resulting in a deep unrolling network with theoretical guarantees. Our extensive numerical experiments demonstrate that our proposed algorithm surpasses existing methods, particularly excelling in fewer scanned views and complex noise conditions.