PROJECT TITLE :

A Novel Retinex-Based Fractional-Order Variational Model for Images With Severely Low Light

ABSTRACT:

A new fractional-order variational model based on Retinex is proposed in this research for low-light photos. Existing integer-order regularisation methods cannot regulate the extent of the proposed method's regularisation. Furthermore, we use a fractional-order gradient total variation regularisation to decompose the image directly in the image domain so that we may achieve more accurate findings. The following are the advantages of the proposed method: The calculated reflectance includes small-magnitude features. Second, the predicted reflectance is free of lighting components. A piecewise smooth estimate of illumination is more probable. We compare the suggested approach with previous Retinex-based approaches that are conceptually similar. The proposed method has been tested and found to be successful.