<p>In image processing, blind deconvolution is a technique designed to simultaneously recover a latent image and the corresponding unknown blur kernel from a degraded or blurred observation. This article presents an innovative framework for blind deconvolution that integrates three key components: bilateral total variation (BTV), Nash game theory, and fractional-order derivatives. Bilateral total variation (BTV) acts as a robust regularizer that is able to keep the edges well and suppress the noise. Another category of methods takes a strategic and joint consideration for the blind deconvolution problem, with the blind deconvolution formulated a two-player game: one player estimates the latent image while the other reestimates the blur kernel. By enclosing complex image structures and considering multiple noise model, fractional-order derivatives can provide more flexibility and accuracy. The proposed method employs an iterative strategy, utilizing the Nash game framework to coordinate the image and kernel estimation tasks. BTV regularization is applied at each step to preserve spatial details, while fractional-order derivatives adapt to variations in degradation patterns. This combination effectively reduces over-smoothing and improves the recovery of fine textures, even in scenarios involving severe blur and non-Gaussian noise. The proposed framework achieves significant improvements in peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) while maintaining computational efficiency. Applications of this method extend to various practical scenarios, including medical imaging, astronomical data analysis, and photographic restoration. This work underscores the potential of integrating advanced regularization, game theory approaches, and fractional calculus in advancing blind deconvolution techniques and image restoration methodologies.</p>