Non-local blind image deblurring approach using game theory

Non-local regularization for image deblurring has been widely studied in recent years. However, minimizing a non-local filter is more difficult than a local one and remains a challenge. The non-local filter can consider not only the local region when changing a particular pixel but also take into account the entire image. In this work, we propose a new non-local method based on total variation (TV) for image deblurring using Nash game theory. This method consists of using two players as a non-local method. Each player has his non-local strategy to find an optimal solution. The first strategy solves the image deblurring problem, while the second uses the point spread function to minimize the non-local function. The experiments show that the new non-local approach gives better results than the local approach and the classical non-local methods, where we have processed four tests that compare our method with others from the literature.