Retouching images is an essential aspect of enhancing the visual appeal ofphotos. Although users often share common aesthetic preferences, theirretouching methods may vary based on their individual preferences. Therefore,there is a need for white-box approaches that produce satisfying results andenable users to conveniently edit their images simultaneously. Recent white-boxretouching methods rely on cascaded global filters that provide image-levelfilter arguments but cannot perform fine-grained retouching. In contrast,colorists typically employ a divide-and-conquer approach, performing a seriesof region-specific fine-grained enhancements when using traditional tools likeDavinci Resolve. We draw on this insight to develop a white-box framework forphoto retouching using parallel region-specific filters, called RSFNet. Ourmodel generates filter arguments (e.g., saturation, contrast, hue) andattention maps of regions for each filter simultaneously. Instead of cascadingfilters, RSFNet employs linear summations of filters, allowing for a morediverse range of filter classes that can be trained more easily. Ourexperiments demonstrate that RSFNet achieves state-of-the-art results, offeringsatisfying aesthetic appeal and increased user convenience for editablewhite-box retouching.