How to Track Body Fat Percentage Without a DEXA Scan

Why tracking body fat matters more than the scale

The bathroom scale tells you one number: how much gravity is pulling on your body. It cannot tell you whether the 2 lbs you lost this week came from fat, muscle, water, or your last meal. Body fat percentage, on the other hand, is the single most useful metric for tracking body composition changes.

When you lose fat and gain muscle at the same time — a process called body recomposition — the scale often doesn't move. Yet your body is visibly changing. Body fat percentage captures that change. Scale weight does not.

The gold standard: DEXA scans

A DEXA (Dual-energy X-ray Absorptiometry) scan uses two low-dose X-ray beams of different energy levels to separately measure bone, lean tissue, and fat tissue. Accuracy is typically within ±1–3% compared to cadaver dissection — the only truly perfect measurement.

The downsides are real:

• Cost: $75–$150 per scan in the U.S.
• Access: You need a clinic or specialized gym
• Frequency: Useful only every 8–12 weeks — you can't DEXA weekly
• Radiation: Minor, but non-zero

For most people tracking a 12–24 week transformation, one DEXA at the start and one at the end is ideal. But you still need a way to track between them.

Home methods ranked by accuracy

1. AI photo-based body scans (±3–5%)

Recent advances in computer vision allow smartphone apps to estimate body fat from a single photo by recognizing muscle definition, adiposity patterns, and skeletal landmarks. TransformAI sends one to four face-blurred photos (front, side, back, legs) together with your height, weight, age and any tape, caliper or smart-scale readings to a vision-language model, and returns an estimate with a confidence range rather than a single number. Photo-only estimates are roughly ±3–5 percentage points; adding a scale reading or measurements tightens the range.

Advantages:

• The TransformAI free tier includes 4 AI scans a month
• Takes about a minute
• Can scan weekly without fatigue
• Pairs naturally with progress photos

Watch out for: same lighting, same pose, same time of day. Variability in photo conditions is the biggest source of error.

2. Skinfold calipers (±3–4% when done correctly)

The classic Jackson-Pollock 3-site or 7-site method uses a caliper to pinch fat at specific anatomical sites (chest, abdomen, thigh for men; triceps, suprailiac, thigh for women). Cheap calipers cost $10–$30.

The catch: technique matters enormously. Self-measurement of your own skinfolds introduces 5–8% error for beginners. Having a partner measure is far better.

3. Tape measurements + Navy formula (±3–4%)

The U.S. Navy body fat formula uses neck, waist, and (for women) hip circumference plus height. It's been validated on over 10,000 service members and is surprisingly accurate for non-athletes.

Men: %BF = 86.010 × log₁₀(waist − neck) − 70.041 × log₁₀(height) + 36.76

Women: %BF = 163.205 × log₁₀(waist + hip − neck) − 97.684 × log₁₀(height) − 78.387

You need a flexible tape measure and consistent measurement sites. This is what the military uses for fitness testing — good enough for Uncle Sam, good enough for tracking progress.

4. Smart scales with bioimpedance (±5–8%)

BIA scales send a tiny electrical current through your body. The speed of the current depends on hydration, muscle mass, and fat. In theory this should give accurate readings.

In practice, hydration fluctuates daily by 2–4 lbs. Your BIA body fat reading will fluctuate 3–5% for the same reason — which is larger than the actual change you're trying to measure. Use BIA only to track trend over weeks, never week-to-week.

5. Progress photos (subjective but invaluable)

Not a number, but the most honest feedback loop. Take a front, side, and back photo every 2 weeks. Same lighting. Same pose. Same underwear.

Scroll back 4 weeks and compare. The mirror adapts to you daily; photos don't.

The stack that actually works

A practical combination that covers each method's blind spots:

1. AI body scan weekly (TransformAI) — trend line you can actually see
2. Tape measurements every 2 weeks — waist, hip, chest, thigh, arm
3. Progress photos every 2 weeks — front/side/back, identical conditions
4. DEXA every 12 weeks (optional) — ground truth check-in

Track all four. When they agree, you can trust the number. When they disagree, trust the tape and the mirror over the scale.

How to take a good AI body scan

The number is only as good as the photo:

• Lighting: Even, natural light if possible. No harsh overhead shadows.
• Pose: Relaxed, arms slightly away from torso, facing the camera straight-on.
• Clothing: Form-fitting or underwear. Baggy clothes break the model.
• Background: Plain wall, no clutter.
• Time of day: Morning, fasted, post-bathroom. Reduces water-weight variance.
• Distance: Full body in frame, camera at chest height.

Standardize these variables and most of the scan-to-scan noise that comes from the photo itself goes away. The estimate is still an estimate — watch the trend over several weeks, not one reading. If you want to go deeper on tape-only and no-equipment methods, see how to check body fat percentage without equipment.

What numbers should you aim for?

Men:

• Essential fat: 2–5%
• Athletes: 6–13%
• Fit: 14–17%
• Average: 18–24%
• Obese: 25%+

Women:

• Essential fat: 10–13%
• Athletes: 14–20%
• Fit: 21–24%
• Average: 25–31%
• Obese: 32%+

Visible abs generally appear around 12% for men, 20% for women — though genetics and fiber distribution matter enormously.

The bottom line

You don't need a DEXA scan to track your body composition. What you need is consistency. Pick a method, stick to the exact same protocol every week, and trust the trend over any individual data point.

TransformAI is currently in TestFlight beta on iPhone. See how it works and ask for a beta invite.