Photo Blur
& Focus Detector

The photo is drawn onto a canvas and converted to grayscale, then two classic focus metrics are computed on the pixels: the variance of the Laplacian (a measure of local contrast between neighboring pixels) and Tenengrad gradient energy (Sobel edge strength). Sharp photos have lots of high-frequency detail, so these values are high; blurred photos smooth everything out, so they drop dramatically. The result is normalized to a 0-100 sharpness score.

No. Everything runs locally in your browser on a canvas element. Your photos are never uploaded, stored, or transmitted anywhere — the tool works even offline after the page loads.

Yes. Drop multiple files or a folder and every photo is scored automatically, then listed worst-first so the blurry ones float to the top. A CSV export includes every file's score and verdict.

The photo is divided into a grid of regions and each one is scored independently. Green regions are sharp, yellow is slightly soft, orange is soft, and red regions are blurry — so you can see that the eyes are in focus but the ears are soft, for example. Low-detail regions (like a plain sky) are marked neutral so they are not mistaken for blur.

The detector uses absolute focus metrics, and the sensitivity control lets you tune the thresholds: 'Standard' suits most photos, 'Strict' flags slightly soft shots for pixel-peepers and product work, and 'Lenient' only flags obviously blurred photos.

No — this is a detector, not a fixer. It tells you which photos are sharp enough to keep, which are worth re-shooting, and where on a single photo focus fell off. For rescue attempts on slightly soft shots, dedicated deconvolution software exists, but nothing truly recreates detail that was never captured.