Composition, not judgment

Explain why the frame feels tiny, cozy, or bold.

Adjust the geometry. See the visual signals. Leave with one edit you can actually try.

Frame controls

Camera height
82attention confidence

Perceived scale

Tiny in context

Low frame occupancy and large surrounding forms make the subject read smaller, while centered placement and moderate contrast keep the mood cozy rather than lost.

Try this crop

Tighten 10%, keep the seat

Bring the frame in evenly while preserving the seat edge as a scale cue.

This model explains composition variables only. It does not infer a person’s body size, identity, health, personality, mood, or intent.

Frame composition is a heuristic

Read the explanation

The source labels occupancy below sixteen very tiny, below twenty five tiny in context, below thirty eight intimate, and larger values dominant. Eighteen percent therefore reads tiny, while forty percent reads dominant. These bars show eighteen and forty percent of a common frame width. This is a composition heuristic rather than body size or identity inference. At occupancy eighteen and horizontal position fifty, the source score is forty eight plus contrast times point two eight, plus sixteen, minus point nine, rounded and capped at ninety six. Contrast sixty four gives eighty one. Contrast twenty eight gives seventy one. The score is labeled attention confidence but is not a calibrated statistical probability. The bars compare the two rounded heuristic outputs. The page stores at most eight saved scenarios locally. Saving a ninth retains the latest eight. These bars compare nine requested saves with eight retained entries. Reset restores the representative sample but preserves saved scenarios. JSON includes inputs, analysis and a crop brief, and copy can fall back to export. It does not infer health, personality or intent.

Reset the workbench?

This restores the representative sample. Saved scenarios stay available.