The 24 Canonical Facial Regions
Unlike generic beauty apps that apply a broad blur filter over the entire face, Dr Maher Vision AI v4.0 reports its findings over 24 distinct anatomical zones with landmark-governed boundaries:
- Upper Forehead (Left, Center, Right): Sun-exposed zones prone to actinic changes and dehydration lines.
- Glabella: The central frown region between the eyebrows.
- Periorbital Skin (Left & Right Upper Eyelids, Under-Eye Infraorbital, Crow’s Feet): Extremely delicate 0.5mm skin vulnerable to dark circles and fine lines.
- Cheeks (Upper Malar, Medial, Lower Buccal on Left & Right): High sebaceous gland density, vascular flushing, and blemish localization.
- Nasal Complex (Bridge & Tip): Prime focal points for sebaceous filaments, shine, and erythema.
- Nasolabial Folds (Left & Right): Dynamic expression lines and localized redness folds.
- Perioral & Lower Face (Upper Lip Skin, Lower Lip Skin, Chin Center, Left & Right Jawline, Submental Neck): Hormonal blemish zones and skin laxity indicators.
The 7 Analyzed Skin Characteristics
Each region is evaluated across six core visible dermatological dimensions:
- 1. Visual Spot Candidates & Blemishes: Multi-scale Hessian blob detection mapped to Hayashi severity bands.
- 2. Facial Redness & Erythema: Illumination-normalized a* chrominance measuring capillary perfusion.
- 3. Pigmentation & Melanin Distribution: Localized L* contrast detecting solar lentigines and uneven tone.
- 4. Skin Surface Texture: Microrelief gradient sharpness and stratum corneum roughness.
- 5. Skin Shine & Sebum Reflection: Specular highlight clustering distinguishing healthy glow from excess sebum.
- 6. Under-Eye Darkness: Infraorbital L* differential measuring periorbital hyperpigmentation and vascular pooling.
Frequently Asked Questions
Why is zero overlap between regions important?
Zero overlap prevents double-counting features and ensures area-weighted aggregation across the face is mathematically sound and reproducible.
Can the 24 regions be customized for different face shapes?
Yes. The 24 regions are dynamically reconstructed for each individual face from 478 MediaPipe landmark coordinates, adapting perfectly to oval, round, square, or heart-shaped facial anatomies.
