PROJECT TITLE :

Matrix Completion Based on Non-Convex Low-Rank Approximation

ABSTRACT:

NNM, a convex relaxation for rank minimization (RM), is a widely used tool for matrix completion and relevant low-rank approximation issues without any prior structural information. Because NNM does not account for the difference between single values, the result it generates is often inconsistent with the solution we had in mind. Non-convex regularisation is used to build two matrix completion models in this research. As a result, we've developed an efficient optimization method with convergence guarantee that's faster than standard approaches at convergence. Furthermore, we show that the suggested regularizer and optimization method may be used to additional RM issues, such as subspace clustering based on low-rank representations. Experimental results on real photos show that the developed models outperform a number of currently available matrix completion techniques by a wide margin. There are also several trials that help us determine how quickly our established optimization strategy converges on a desired solution.