Visual Detection of Facial Spots
Facial blemishes present distinct photometric signatures: localized circular depressions or elevations characterized by increased a* redness (erythema) or decreased L* luminance.
Dr Maher Vision AI v4.0 employs multi-scale Difference-of-Gaussians (DoG) filtering combined with Hessian matrix curvature checks. A valid spot candidate must exhibit balanced eigenvalues (λ₁/λ₂ ≈ 1), preventing linear wrinkles or shadow folds from being misclassified as blemishes.
The Hayashi Severity Scale
In clinical dermatology literature (Hayashi et al.), facial inflammatory lesions are categorized into standardized count brackets (0: none, 1–5: mild, 6–20: moderate, 21–50: severe, >50: very severe). Our vision engine reports visual candidate counts mapped to these established bands for reproducible interpretation.
Frequently Asked Questions
Can the scan distinguish acne from freckles or moles?
The engine uses chromatic thresholds (a* redness vs melanin absorption) to differentiate reddish inflammatory marks from brown pigmented lentigines, but dermatoscopic evaluation is necessary for precise lesion classification.