The science

Your targets aren't magic —
they're math you can check.

Every number CalorieLensAI shows you — daily calories, protein grams, fibre goals — comes from published equations that dietitians and sports scientists have used for decades. Here's exactly what we compute, why, and how it helps you actually stick to your goals.

See the formulas

What's in this page

  1. Why food tracking actually works
  2. The formulas we use
  3. 1. Resting energy (BMR — Mifflin-St Jeor)
  4. 2. Daily energy (TDEE — activity multiplier)
  5. 3. Deficit or surplus for your goal
  6. 4. Protein target
  7. 5. Fat & carbohydrate split
  8. 6. Fibre target
  9. A worked example, end to end
  10. How the AI estimates portions
  11. How this helps you stay on track
  12. References

Why food tracking actually works

Randomised trials and large observational studies consistently find that the single strongest behaviour predictor of weight-loss and weight-maintenance success is self-monitoring — writing down what you ate, usually the day you ate it. Reviewing published data:

  • In a 6-month behavioural weight-loss trial (Kaiser Permanente, Am J Prev Med 2008), participants who kept daily food records lost roughly twice as much weight as those who kept none.[1]
  • A 2011 meta-review in the Journal of the American Dietetic Association concluded self-monitoring shows a consistent, significant association with weight-loss outcomes across 22 studies.[2]
  • Even in maintenance, the National Weight Control Registry finds that people who kept the weight off for >5 years weigh themselves and track intake more consistently than the general population.[3]
i

The insight: awareness closes the gap between what you think you ate and what you actually ate. Photo-first tracking eliminates the biggest reason people quit — the friction of typing.

The formulas we use

When you complete onboarding, CalorieLensAI computes five numbers: your target calories, protein, carbohydrates, fat and fibre. Each one comes from an established equation. No black boxes.

1. Resting energy — BMR (Mifflin-St Jeor)

Basal Metabolic Rate is the energy your body burns at rest — just keeping cells alive, breathing, and thinking. Roughly 60–75% of your total daily calories go here.

We use the Mifflin-St Jeor equation (1990), which multiple reviews — including the American Dietetic Association's 2005 evidence analysis — have shown is the most accurate predictor of resting energy expenditure in non-obese and obese adults.[4][5]

Mifflin-St Jeor · male
BMR = 10 × weight (kg) + 6.25 × height (cm) − 5 × age (yr) + 5

Mifflin-St Jeor · female / other
BMR = 10 × weight (kg) + 6.25 × height (cm) − 5 × age (yr) − 161

2. Daily energy — TDEE (activity multiplier)

You burn more than just BMR when you walk, work, and move. To turn BMR into Total Daily Energy Expenditure we multiply by a published activity factor. The values below are the Harris-Benedict / FAO/WHO/UNU activity factors used across sports-nutrition literature.[6]

Activity levelWhat it looks likeMultiplier
SedentaryDesk job, little intentional exercise× 1.20
LightLight exercise 1–3× / week× 1.375
ModerateExercise 3–5× / week× 1.55
ActiveHard exercise 6–7× / week× 1.725
Very activePhysical job + daily training× 1.90
TDEE = BMR × activity multiplier

3. Deficit or surplus for your goal

To lose fat we need an energy deficit; to build muscle we need a modest surplus. The magnitudes below are drawn from the ISSN and Academy of Nutrition & Dietetics position stands and are chosen to be sustainable — aggressive cuts cause rebound, aggressive bulks cause fat gain.[7][8]

GoalAdjustmentWhy
Lose weightTDEE × 0.80≈20% deficit ≈ 0.5–0.75 kg / wk loss
Maintain / Improve healthTDEE × 1.00Zero net energy change
Gain muscleTDEE × 1.15≈15% surplus supports lean mass without excess fat

We enforce a hard floor at 1,200 kcal/day — below that adequate micronutrient intake is very hard to achieve and adherence collapses. If your calculated deficit would go lower, we cap it. This matches the Academy of Nutrition & Dietetics' safe-minimum guidance for adult women, and is a common ceiling used in commercial programs.[9]

4. Protein target

Protein does more than build muscle: it's the most satiating macronutrient (highest thermic effect + slowest gastric emptying), and preserves lean mass during a caloric deficit. The International Society of Sports Nutrition position stand recommends 1.4–2.0 g protein / kg bodyweight for active individuals; the higher end during a deficit or gain phase.[10]

Default target
protein_g = weight (kg) × 1.6

Gain phase (extra emphasis on lean mass)
protein_g = weight (kg) × 1.8

# You can override this manually during onboarding
✓

Why we default to 1.6 g/kg not 0.8 g/kg: The RDA of 0.8 g/kg is a minimum-to-avoid-deficiency value for sedentary adults. Every controlled study looking at satiety, body composition or muscle retention lands substantially higher — typically 1.6 g/kg or above.

5. Fat & carbohydrate split

After protein is set, we allocate 28% of calories to fat and let carbohydrates fill the remaining energy budget.

  • Fat 28% — comfortably inside the U.S. Institute of Medicine's Acceptable Macronutrient Distribution Range of 20–35%, high enough to support hormone production, low enough to leave carbs for training energy.[11]
  • Carbs = whatever's left, with a floor of 50 g/day so glycogen and brain function stay covered. Fibre is included in this number and set explicitly (below).
fat_g = calories × 0.28 ÷ 9  # 9 kcal per gram of fat
carbs_g = (calories − protein_g × 4 − fat_g × 9) ÷ 4  # 4 kcal per gram
then carbs_g = max(50, carbs_g)

6. Fibre target

The U.S. Institute of Medicine Dietary Reference Intake for fibre is 14 g per 1,000 kcal — a single, calorie-scaled number that lands on ~25 g/day for women and ~38 g/day for men eating a standard 2,000–2,700 kcal.[11] This is the value we use.

fiber_g = calories ÷ 1000 × 14  # floored at 20 g/day

Adequate fibre is associated with lower all-cause mortality, better blood-sugar control, and — relevant to weight loss — meaningfully improved satiety on the same calorie budget.

A worked example, end to end

Meet Ethan: 32 years old, 78 kg, 180 cm, moderately active, wants to lose weight.

Ethan's daily targets

BMR (Mifflin-St Jeor · male)10·78 + 6.25·180 − 5·32 + 5 = 1,750 kcal
TDEE (× 1.55 moderate)1,750 × 1.55 ≈ 2,713 kcal
Goal-adjusted (× 0.80 lose)2,713 × 0.80 ≈ 2,170 kcal
Protein (78 × 1.6)125 g  ·  500 kcal
Fat (2,170 × 0.28 ÷ 9)68 g  ·  612 kcal
Carbs (remainder ÷ 4)264 g  ·  1,056 kcal
Fibre (2,170 ÷ 1000 × 14)30 g
Total2,168 kcal — checks out ✓

Ethan sees these five numbers on his dashboard; every meal he snaps subtracts from them in real time.

How the AI estimates portions from a photo

Once your daily targets are set, the second half of the equation is knowing what you actually ate. This is where the photo-first approach earns its keep.

The two-stage pipeline

  • Vision model. Every photo is sent to GPT-4o Vision (via our secure API). It identifies each dish, its cooking method, and estimates portion size in natural units (grams, slices, cups, oz).
  • Nutrition mapping. Detected foods are matched against a nutrition database. Values return with a per-item confidence score (0–100%) so you know what to double-check.

Why editing matters

The AI is a starting point, not a verdict. Every value is editable in natural units — "3 eggs" not "168 g" — and totals recalculate instantly. Reviews of consumer food-tracking apps have shown that human-in-the-loop review closes almost all of the accuracy gap against gold-standard weighed-food records. We designed the UX to make that review take about 5 seconds.

!

Honest limits: AI estimates are typically within ±10–20% for common dishes with reasonable lighting. Mixed casseroles, sauces, and unusual portions are harder. That's why every item is editable and confidence-scored.

How this actually helps you stay on your goals

The math above is only useful if the daily behaviour keeps happening. We designed CalorieLensAI around five findings from the behaviour-change literature:

1. Reduce the effort of self-monitoring

The single biggest predictor of tracking-app abandonment is entry friction. Photo entry cuts the median time from ~45 s (search-and-type) to ~8 s. Lower friction → higher adherence → better outcomes.[12]

2. Show progress in trends, not one-day totals

Body weight and adherence oscillate day-to-day; motivation should not. Our Trends screen surfaces 7- and 30-day averages so a bad Tuesday doesn't tank your Monday.

3. Non-punitive design

Streaks that support, not shame. No red numbers. No "you failed today" language. Reviews of eating-disorder outcomes consistently flag punitive-tracking UX as a risk factor — we designed against it.

4. Editable, transparent numbers

Every formula on this page is visible to you. Every macro on every meal is editable. Trust in the numbers builds when nothing is hidden.

5. Protein-forward defaults

Because we default protein to 1.6 g/kg rather than the RDA 0.8 g/kg, users hitting their target feel full at the calorie budget instead of hungry. Sustained satiety is the strongest lever against overshooting.

Ready to build your targets?

Two minutes of onboarding. The math runs live in front of you.

Get started free

References

  1. Hollis JF, Gullion CM, Stevens VJ, et al. Weight loss during the intensive intervention phase of the weight-loss maintenance trial. Am J Prev Med. 2008;35(2):118-126.
  2. Burke LE, Wang J, Sevick MA. Self-monitoring in weight loss: a systematic review of the literature. J Am Diet Assoc. 2011;111(1):92-102.
  3. Klem ML, Wing RR, McGuire MT, et al. A descriptive study of individuals successful at long-term maintenance. National Weight Control Registry. Am J Clin Nutr. 1997;66(2):239-246.
  4. Mifflin MD, St Jeor ST, Hill LA, et al. A new predictive equation for resting energy expenditure in healthy individuals. Am J Clin Nutr. 1990;51(2):241-247.
  5. Frankenfield D, Roth-Yousey L, Compher C. Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic review. J Am Diet Assoc. 2005;105(5):775-789.
  6. FAO/WHO/UNU. Human energy requirements. FAO Food and Nutrition Technical Report Series 1. Rome, 2004.
  7. Aragon AA, Schoenfeld BJ, Wildman R, et al. International Society of Sports Nutrition position stand: diets and body composition. J Int Soc Sports Nutr. 2017;14:16.
  8. Academy of Nutrition and Dietetics. Position of the Academy of Nutrition and Dietetics: interventions for the treatment of overweight and obesity in adults. J Acad Nutr Diet. 2016;116(1):129-147.
  9. American College of Sports Medicine, Academy of Nutrition and Dietetics, Dietitians of Canada. Nutrition and Athletic Performance. Joint Position Statement. Med Sci Sports Exerc. 2016;48(3):543-568.
  10. Jäger R, Kerksick CM, Campbell BI, et al. International Society of Sports Nutrition Position Stand: protein and exercise. J Int Soc Sports Nutr. 2017;14:20.
  11. Institute of Medicine (US). Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. National Academies Press. Washington DC, 2005.
  12. Cordeiro F, Bales E, Cherry E, Fogarty J. Rethinking the mobile food journal. Proc CHI Conf Human Factors in Comp Systems. 2015:3207-3216.