Food Quantity Detection

Learn more about image preprocessing and segmentation in the Image Pre-processing Tutorial page.

Plan & Token Requirements

Feature available in the following LogMeal Plans:

Accessible by the following User Types:

What It Does

The Quantity Estimation feature computes the amount of food (in grams or milliliters) for each detected dish or item in a meal image or image sequence. It uses advanced computer vision and depth modeling to estimate food portions directly from RGB or RGB+Depth captures, without requiring physical reference objects.

This enables accurate energy (kcal) and nutrient quantification when combined with Dish Recognition and Nutritional Analysis, supporting automated dietary logging and nutritional tracking.

✕

✕

When to Use It / Outcomes

• When you need to automatically estimate serving sizes from a photo or multi-angle capture.
• When you want to quantify actual consumption without requiring user input.
• When you need volume and weight estimation for dishes in food diaries, wellness, or nutrition tracking apps.

Feature-Specific Details

• Inputs: image sequence captured around the meal (via Depth SDK).
• Outputs: numerical estimates for weight (grams) for each detected food item.
• Integration: Designed to work seamlessly with Dish Recognition and Ingredients Detection.
• Units: Metric (g, ml) by default; optionally convertible to imperial units (oz, cups).
• Personalization: Takes into account camera distance, plate geometry, and image perspective for improved precision.
• Advanced Usage: For highest accuracy, use with the LogMeal Depth SDK described in the Food Quantity Detection use case.

Related Endpoints

Use the following endpoints for quantity estimation:

• POST /image/segmentation/complete/quantity → 🔴 Estimates quantity for detected dishes in a captured sequence.

Related Use Cases