Block 01 · C1 · +64 px
Confidence 0.93. The screen's thin top-left protruding rail and its cyan/red fringe end inside +64, leaving a narrow but clean residual gap.
DiagSplat · local multiradius boundary critic
每个 block 的 RGB 和 residual 均作为独立 512×512 PNG 输入。 C1 白=+64,C2 洋红=+128,C3 黄=+192,C4 薄荷绿=+256 px。 选择最小的完整覆盖候选后,以 Gaussian 空间权重平滑组合局部半径,避免方形 crop 接缝。
Residual 仍是有偏辅助:轮廓外 residual 可以否决候选, 但低 residual 不能证明覆盖。最终组合使用 RGB-first 的选择。
Confidence 0.93. The screen's thin top-left protruding rail and its cyan/red fringe end inside +64, leaving a narrow but clean residual gap.
Confidence 0.94. The full horizontal top rail and small left tip are contained below the white line; no coherent screen trail crosses +64.
Confidence 0.91. The left upright, foot, and broad cyan defocus plume remain to the right of +64; the plume fades before the contour.
Confidence 0.91. At this alternate left-side position, the screen edge and diffuse color trail again terminate inside the white boundary, though with less margin than the larger candidates.
Hard mask fraction 18.181% → 21.061%;新增 2.880%。 下载最终 mask PNG
Judge ONE frame-0022 boundary location at a time. You receive two separate 512x512 PNG images for the same location: A. true observed RGB; B. biased Round-1 residual. They are separate visual inputs, never a merged composite. Four thin candidate hard-mask boundaries appear in both images: C1 white/black = +64 px C2 magenta/black = +128 px C3 yellow/black = +192 px C4 mint/black = +256 px Choose the SMALLEST candidate that encloses every visible part of the local transient object indicated by the current mask. Include thin tips, small protrusions, translucent or defocused boundaries, motion/defocus halos, and color/blur trails. The existing mask may under-cover the object. Do not expand merely to include unrelated static background. If even C4 fails, return INSUFFICIENT. Residual construction: mean absolute RGB difference between dirty observed and the Round-1 3DGS render, Gaussian sigma 1.8, clipped at 0.30, Turbo colormap. The render was trained with an earlier undercovered mask and can reproduce leaked transient pixels. Therefore coherent residual outside a line proves that line insufficient, but low residual never proves coverage. Return: selected candidate, radius, confidence, and one concise RGB-first reason.