Types of Face Shapes: All 7 Explained, With How to Test

Introduction

There are seven recognized face shapes: oval, round, square, rectangle (oblong), heart, diamond, and triangle. They’re classified by comparing measurements of your forehead, cheekbones, jaw, and face length against each other. This guide covers all seven, the science behind the classification, how to test yourself, and why different tools and quizzes often disagree.

What Is Face Shape?

Face shape is the general outline your face makes when you remove hair, expression, and styling from the picture — just the outer boundary formed by your forehead, cheekbones, jaw, and chin. It’s a simplification. Real faces are three-dimensional and don’t sort perfectly into seven clean buckets, but the classification is useful shorthand for describing proportions in a way that’s consistent from one person to the next.

Face Shape vs. Face Structure

These two terms get used interchangeably, but they’re not quite the same thing. Face structure refers to the full underlying bone architecture — brow ridge projection, cheekbone height, jaw angle, chin projection. Face shape is the flattened, two-dimensional outline that results from that structure.

Two people can share the same face shape classification while having noticeably different face structures underneath — one might have a stronger jaw angle, the other softer, but if both taper similarly from cheekbone to chin, they’ll both read as “heart” or “oval.”

Shape is what a quick visual classification captures; structure is the fuller picture a trained eye or a 3D scan would use.

Why Face Shape Matters

Face shape is mainly useful for two practical decisions: where to add or reduce volume in a haircut, and which glasses frame shapes will contrast well rather than repeat your face’s own lines. It’s a styling reference point, not a measure of attractiveness, health, or anything beyond geometry.

The Scientific Basis: Anthropometry

Face shape classification borrows directly from craniofacial anthropometry — the branch of physical anthropology that measures facial proportions. A few terms come up often in this research:

• Bizygomatic width — the distance between the two zygion points (the widest part of the cheekbones). This is typically around 60–75% of vertical face height and is the single most-used width measurement in facial classification.
• Facial width-to-height ratio (fWHR) — bizygomatic width divided by the distance from the brow line to the upper lip. It’s a well-studied ratio in facial research, though scientists have flagged that ratio-based measurements like this can produce statistically misleading patterns if not adjusted for overall face size (a limitation worth knowing if you see fWHR cited as more definitive than it is).
• Mandibular (bigonial) width — the distance between the two widest points of the jaw, usually around 70–75% of bizygomatic width. This is what separates square/rectangle (jaw close to cheekbone width) from heart or triangle (jaw clearly narrower or wider than the cheekbones).
• Forehead (bitemporal) width — typically 80–85% of bizygomatic width. Where this sits relative to jaw width is what separates heart (wide forehead) from triangle (narrow forehead).
• Craniofacial landmarks — the specific reference points (zygion, gonion, trichion, menton, and others) that anthropometric and AI systems both measure from. These fixed points are what make the classification consistent from one face to the next, rather than a subjective eyeballing exercise.

These aren’t arbitrary categories — they’re the same measurement points used in decades of physical anthropology research, adapted for styling rather than clinical or forensic use.

How Face Shape Is Determined: Classification Methods?

There are two practical ways to classify a face shape, and they don’t always agree.

Manual measurement.

Using a tape measure or string, you record forehead width, cheekbone width, jaw width, and face length, then compare the ratios by hand.

AI landmark detection.

A photo-based tool maps anywhere from 68 to 478 points on the face and calculates the same ratios computationally, using computer vision models built on the same anthropometric principles.

A note on limitations:

AI facial analysis tools give styling guidance based on geometry — they’re not medical or diagnostic tools, and their output should be treated as an estimate rather than a certified measurement.

Comparison Table: The 7 Face Shapes

Quick Diagnostic: What Shape Am I?

Is your face noticeably longer than it is wide?
├── No → Are forehead, cheeks, and jaw close to equal width?
│         ├── Yes → Is your jaw angular?
│         │         ├── Yes → Square
│         │         └── No  → Round
│         └── No  → Is your forehead the widest point?
│                   ├── Yes → Heart
│                   └── No  → Is your jaw the widest point?
│                             ├── Yes → Triangle
│                             └── No  → Diamond
└── Yes → Are cheekbones the widest point, with a soft jaw curve?
          ├── Yes → Oval
          └── No  → Rectangle (Oblong)

This won’t replace a proper measurement, but it narrows things down fast if you already know your basic proportions.

Common Misclassifications

• Oval vs. rectangle/oblong — the most frequent mix-up. Both are longer than wide with cheekbones/even width as the widest area; the difference is that oval curves inward toward a rounded chin, while rectangle keeps straight sides down to an angular jaw.
• Diamond vs. heart — a forehead that’s only slightly wider than the jaw can get read as either, depending on the classifier or measurer’s judgment call.
• Square vs. rectangle — often confused because both have angular jaws and even width; face length is what separates them.

AI vs. Manual Measurement

Manual measurement is free and requires no photo, but it’s sensitive to where exactly you place the tape, hair covering the hairline, and measuring at a slight angle. AI landmark detection removes most of that human error by locking onto fixed facial landmarks from a still photo, which is why two people measuring themselves by hand sometimes land on a different result than a photo-based tool gives them.

Neither method is flawless — AI results shift with lighting, expression, and camera angle too — but landmark-based detection is generally the more consistent of the two.

Does Face Shape Change? Genetics, Aging, and Weight

Genetics set your underlying bone structure — the position of your cheekbones, jaw angle, and brow ridge are established by adulthood and don’t change afterward. Aging typically reduces cheek fullness and jaw definition over time, which can shift someone from a rounder or oval classification toward a more angular or oblong one later in life.

Weight change affects the soft tissue around the bone structure — significant weight gain or loss can move a border-line face (say, round/oval) toward the other side of that border, without changing the bone underneath. Certain cosmetic procedures (jaw-slimming injections, buccal fat removal) can produce a similar shift by altering soft tissue volume rather than bone.

A note on ethnicity: bone structure proportions do vary somewhat across populations, which is one reason a single global percentage breakdown of face shapes doesn’t fully hold — see the statistics section below.

Face Shape Statistics: Why the Percentages Never Agree

You’ll see very different numbers depending on the source — some report oval at roughly 28% of people, others as high as 46%. A few things explain the gap:

• Sample bias — most published percentages come from usage data on a specific site’s own face-shape tool, not a random population survey. People who seek out a face-shape detector aren’t a neutral cross-section.
• Dataset size and scope — larger datasets aren’t automatically more representative if they’re still drawn from the same self-selected pool of app users.
• Classifier differences — tools disagree on the exact ratio cutoffs for each shape (for example, whether “oval” starts at a 1.25 or a 1.5 length-to-width ratio), so the same face can be classified differently by two tools.
• Cultural and population variation — average facial proportions differ somewhat across ethnic populations, so a percentage breakdown drawn from one dataset won’t necessarily generalize globally.

What holds up across most independent sources, even when exact numbers move, is the ranking: oval, round, and oblong are consistently reported among the most common results, and diamond consistently ranks among the rarest.

Treat any single specific percentage — including ones cited elsewhere on this topic — as directional rather than exact.

Face Shape Myths

• “Diamond is the rarest, so it’s the most desirable.” Rarity and attractiveness aren’t linked — this is a common but unsupported leap.
• “Your face shape is fixed forever.” Not entirely true — see the aging and weight section above.
• “Oval is the objectively best/most attractive shape.” Oval is the most versatile for styling because it’s close to symmetrical proportions, which is different from being the most attractive. All shapes support flattering styling choices.
• “A quiz result is as accurate as a measurement.” Subjective quizzes (“does your jaw look angular?”) are meaningfully less reliable than tape measurement or AI landmark detection.

FAQs

Is it 6, 7, or 8 face shapes?

Most current guides and AI detectors use 7. Older 6-shape guides usually merge rectangle and oblong into one category; some 8-shape guides split triangle into standard and inverted versions.

Which face shape is the rarest?

Diamond is the shape most consistently reported as rarest across independent sources, though exact percentages vary by dataset.

Which face shape is considered most attractive?

There’s no settled answer — attractiveness research doesn’t map neatly onto the 7-shape styling classification, and preferences vary by culture and era. Oval is the most versatile, which is a different claim than being the most attractive.

Can surgery or braces change your face shape?

Surgery that alters bone or significant soft tissue (jaw contouring, buccal fat removal) can shift your classification. Braces primarily move teeth and can subtly affect jaw appearance over long treatment but don’t change your overall face shape category.

Does losing weight change your face shape?

It can shift borderline cases — most often between round and oval — by reducing cheek and jaw soft tissue, without changing the bone structure underneath.

Does face shape affect hairstyle choice?

Yes — it’s one of the two areas (along with glasses) where face shape guidance is most consistently useful, mainly for deciding where to place volume.

Does face shape change with age?

Often, yes. Reduced cheek fullness and jaw definition with age can shift someone toward a more angular or elongated classification later in life.

Can men and women have the same face shape?

Yes — all seven shapes occur in both sexes, though some shapes (heart, for instance) are reported more frequently in women, and others (square, triangle) more frequently in men.

Why do AI face shape tools disagree with each other?

Different tools use different landmark models and different ratio cutoffs for each shape category, so the same face can be classified differently depending on which tool measures it.

Which face shape suits the widest range of hairstyles?

Oval, because its proportions are already close to balanced, which is why it’s typically used as the reference point when designing hairstyle guidance for other shapes.

Conclusion

Face shape is a spectrum, not a fixed label — most people sit close to a boundary between two categories rather than squarely inside one. AI landmark analysis is generally more consistent than manual measurement, but neither is exact. And every styling recommendation tied to face shape is a starting guide, not a strict rule.