Problem
Subject silhouette contains a narrow gap that closes after downscaling and changes the perceived shape.
Solution
Root Cause / Diagnostic:
Narrow gaps in subject poses—such as a small negative space between an arm and a torso or between fingers—collapse when downsampled. Bicubic and bilinear scaling algorithms blend adjacent pixels, fusing the narrow gap into a solid mass. This turns a dynamic gesture into an awkward, bulky silhouette that looks deformed on mobile.
Actionable Fix:
1. Pose subjects with decisive, broad silhouettes, ensuring negative space channels between limbs and torso are at least 40px wide.
2. Manually retouch or composite the cutout mask to cleanly open up or decisively close ambiguous micro-gaps.
3. Inspect downscaled silhouettes at 120px width in pure black-and-white to confirm graceful, readable body geometry.
Pro Tip:
A readable silhouette makes or breaks mobile recognition; pose with broad, open gestures so body contours remain distinct at small scales.
Narrow gaps in subject poses—such as a small negative space between an arm and a torso or between fingers—collapse when downsampled. Bicubic and bilinear scaling algorithms blend adjacent pixels, fusing the narrow gap into a solid mass. This turns a dynamic gesture into an awkward, bulky silhouette that looks deformed on mobile.
Actionable Fix:
1. Pose subjects with decisive, broad silhouettes, ensuring negative space channels between limbs and torso are at least 40px wide.
2. Manually retouch or composite the cutout mask to cleanly open up or decisively close ambiguous micro-gaps.
3. Inspect downscaled silhouettes at 120px width in pure black-and-white to confirm graceful, readable body geometry.
Pro Tip:
A readable silhouette makes or breaks mobile recognition; pose with broad, open gestures so body contours remain distinct at small scales.