SECOI: an application based on fuzzy soft sets for producing selective-colored images

We introduce SECOI, an application that creates a selective-colored image. The application is fast and works without adverse effects caused by noisy pixels. It is based on the theory of fuzzy soft sets by evaluating the relation that two colors are similar and; moreover, there are other pixels with the same similar color in their vicinity. The color filtering is then realized by the proper selection of α -cut of the set. We propose SECOI as a global method or as SECOI+dist, where by involving spatial weighting, we create a local method with increased performance. According to the benchmark, SECOI+dist is the most precise method among the tested, and considering Dice score, it yields nearly half error than the runner-up.