Food Recognition Capabilities

The LogMeal API provides AI-powered food recognition at multiple semantic levels.
This means you can analyze a food image and extract different layers of information — from simply detecting if there is food, to recognizing specific dishes, ingredients, and nutritional information.

---

What We Recognize

LogMeal can detect and classify food in several ways. From more general to more specific:

1. Food Type Detection
   Detects whether an image contains prepared food, drinks, fresh fruits/vegetables, sauces, or even if it is not food at all.
   → Useful for large-scale image categorization.

2. Food Group Detection
   Identifies broad categories like meat, fish, vegetables, eggs, bread, legumes, desserts, dairy, and more.
   → Ideal for diet analysis or nutritional grouping.

3. Dish Recognition
   Recognizes 1300+ dishes from international cuisines.
   The list is constantly growing as new dishes are added to the recognition models.
   → Example: distinguish “spaghetti bolognese” vs. “caesar salad.”
   → You can retrieve the complete list of currently supported dishes via GET /dataset/dishes

4. Ingredients Detection
   Extracts ingredients present in dishes, including estimated quantities.
   → Enables detailed nutritional analysis and recipe tracking.

5. Nutritional Information
   Provides calories, macros, and micronutrients based on detected dishes and ingredients.
   → Perfect for diet apps and nutrition tracking.

---

Advanced Capabilities

With higher-tier plans, additional AI features become available:

• Quantity Estimation (Recommend or higher)
  Estimates food volume from RGB or depth images.
  → Supports serving size detection and precise nutrition logging.

• Food Waste Detection (Recommend or higher)
  Compares pre-meal and post-meal photos to detect leftovers.
  → Useful for sustainability and food waste reduction.

• Variety Score & Nutri-Scores (Recommend or higher)
  Computes diversity of diet over time and standardized nutrition quality scores (A–E, 0–100).

• Recipe & Dish Recommendations (Recommend or higher)
  Suggests alternative dishes or recipes based on user preferences, restrictions, and history.

---

Plans & Access

Food recognition is available in all plans, but the level of detail and advanced features vary:

---

Datasets

The LogMeal API provides endpoints to retrieve the full lists of detectable items:

• GET /dataset/dishes → Returns the list of supported dishes
• GET /dataset/ingredients → Returns the list of supported ingredients
• GET /info/nutrients → Returns the list of supported nutritional indicators

---

Related Endpoints

Here are some of the key endpoints for food recognition:

• POST /v2/image/segmentation/complete — detect multiple food items on an image and recognize them and log a user intake.
• POST /v2/nutrition/recipe/ingredients — detect the list of ingredients and their quantities for an intake.
• POST /v2/nutrition/recipe/nutritionalInfo — detect the nutritional information for an intake (energy, carbs, protein, fats, etc. 35+ nutritional indicators).
• GET /v2/history/getIntakesList — display a food diary, retrieve the list of intakes performed by a user in a period of time.

More in the API Documentation explorer.

---

Next Steps:

• Learn about Plans & Limits
• See how to Preprocess Images for Better Recognition