Katharine Moon Vari, RISD Mellon Intern 2026
This Gramcko piece currently on display in the galleries stood out to me because it has many markers of traditional figure-on-ground: the dark black and red stand out on the light yellow and white. Despite the lack of resemblance of the shape to anything in real life, the placement of the white below the yellow feels similar to a floor before a background wall. I found that the way the black section outlined the red made them appear to be figures on the same plane. I wasn't surprised to see how well the SAM model was able to find masks-the clear lines and separation of elements by color makes delineating elements quite clear.
Working with this Newman piece was difficult. I was confident in choosing elements that I felt stuck out in space, but I was aware of many alternate choices and am curious to see if other viewers would interpret the space differently. When I saw the elements in a visualization like this, the piece felt so different from seeing the original two-dimensional work on my screen. It started to look like a folded out cube, which I never would have thought of without seeing it in this way.
This visualization of Minako Iwamura's Sixty Five demonstrated to me some of the limitations of the computer algorithm. Because of its inability to recognize color indicators of transparency (due to its intended usage in masking), I felt like some of the elements of the painting were lost. Also, the SAM made strange decisions about the borders with sharper corners. However, I was surprised by how fun the resulting visualization was, especially its emphasis on colored gradients in the background. This was a fun one to make!