A Color-Based Selective and Interactive Filter Using Weighted TV

In this paper we propose to introduce a new color interactive and selective filtering tool based on the minimization of a weighted vectorial total variation term TVg with the L2 norm as data term. Our goal is to filter one region of an image while preserving the other using a minimal interaction with the user. To this end, we take benefit of a weighted vectorial regularization term TVg based on color moments in order to perform our selective color filtering. Up to now, color moments have been mainly introduced for indexation purposes. In our case, the user selects some points in the area to preserve and some other points in the area to be filtered. Reference color moments are then computed on patches around the selected points and are included in the TVg term through various functions g. Two main g functions are tested within the HSL color space leading to very interesting results on both synthetic and real images. Convex optimization tools are then used to solve the minimization issue. An augmented Lagrangian formulation leads to a simple and efficient algorithm based on Uzawa block relaxation schemes. Our algorithm, while easy to be implemented, proves to be efficient in terms of computational cost due to its robustness towards the choice of the penalty parameter. The proposed filtering tool may be very interesting as a pre-processing step for segmentation, movie post-production or object-oriented compression.