Method guide

How local photo sharpness analysis works

Sharpest Shot Finder does not run a remote AI service. It decodes each selected image locally, makes a bounded analysis copy, and sends pixel data to a small Web Worker pool.

Sharpness signals

The technical sharpness score combines variance of the Laplacian, edge density, local contrast and directional gradient imbalance. These signals react differently to soft focus, general blur and one-direction motion smear.

Exposure signals

Highlight clipping counts nearly white pixels and shadow clipping counts nearly black pixels. The exposure score also considers how far the overall brightness is from the middle of the range. It is a technical warning, not an aesthetic rule.

Near-duplicate grouping

A 64-bit difference hash represents broad luminance structure. Images with matching dimensions, small hash distance and nearby timestamps are grouped. The sensitivity control changes both the hash and time thresholds.

Human review

“Sharpest in this group” means the highest technical score among the grouped files. It does not inspect expression, emotional value, composition or the exact subject you care about.