Granted Patent B2

US 12,394,222 · App. 18/090,391 · Granted Aug 19, 2025

Inventors: Dae Hoon Kim (Seoul, KR); Jey Yoon Ru (Seoul, KR); Seung Woo Ji (Seoul, KR)

Assignee: NUVI LABS CO., LTD.

G06V20/68G06F16/583G06T7/11G06V10/82G06V20/70G06T2207/30128

Abstract

Claims (23)

1. A method for analyzing food executed by an apparatus for analyzing food, the method comprising:

training an image encoder, a second text encoder, and a third text encoder using contrastive learning using a contrastive language-image pre-learning structure;

generating image captioning data using food image features extracted from a food image,

wherein the generating image captioning data extracts a first embedding having food image features through the image encoder to which the food image is input and generates image captioning data including food ingredients for the food image by inputting the extracted first embedding to the first text decoder, and wherein the generating image captioning data extracts a second embedding having food ingredient features through the second text encoder to which the inferred food ingredients are input and generates image captioning data including food recipes for the food image by combining the first embedding having the extracted food image features and the extracted second embedding and inputting the combination to the second text decoder; and

generating a food name for the food image using the generated image captioning data,

wherein the generating a food name extracts a third embedding having food recipe features through a third text encoder to which the generated food recipe is input and generates a food name for the food image by combining the extracted first, second, and third embeddings and inputting the combination to the third text decoder.

2. The method of claim 1 , wherein the generating image captioning data generates image captioning data including food ingredients using food image features extracted from the food image.

3. The method of claim 1 , wherein the generating image captioning data infers food ingredients using food image features extracted from the food image and generates image captioning data including food recipes for the food image using the inferred food ingredients.

4. The method of claim 1 , wherein the image encoder, the second text encoder, and the third text encoder are learned so that a triplet of the first to third embeddings for the same food image, food ingredient, and food recipe is more similar to each other than a triplet of embeddings for different food images, food ingredients, and food recipes.

5. Method of claim 3 , further including analyzing food nutrients in the food image using at least one of the inferred food ingredient, the generated food recipe, and the generated food name.

6. An apparatus for analyzing food using image captioning comprising:

a memory storing one or more programs; and

a processor executing the one or more programs stored,

wherein the processor is configured to:

train an image encoder, a second text encoder, and a third text encoder using contrastive learning using a contrastive language-image pre-learning structure;

generate image captioning data using food image features extracted from a food image,

wherein the processor extracts a first embedding having food image features through the image encoder to which the food image is input and generates image captioning data including food ingredients for the food image by inputting the extracted first embedding to the first text decoder, and wherein the processor extracts a second embedding having food ingredient features through the second text encoder to which the inferred food ingredients are input and generates image captioning data including food recipes for the food image by combining the first embedding having the extracted food image features and the extracted second embedding and inputting the combination to the second text decoder; and

generate a food name for the food image using the generated image captioning data,

wherein the processor extracts a third embedding having food recipe features through the third text encoder to which the generated food recipe is input and generates a food name for the food image by combining the extracted first, second, and third embeddings and inputting the combination to the third text decoder.

7. The apparatus of claim 6 , wherein the processor generates image captioning data including food ingredients using food image features extracted from the food image.

8. The apparatus of claim 6 , wherein the processor infers food ingredients using food image features extracted from the food image and generates image captioning data including food recipes for the food image using the inferred food ingredients.

9. The apparatus of claim 6 , wherein the image encoder, the second text encoder, and the third text encoder are learned so that a triplet of the first to third embeddings for the same food image, food ingredient, and food recipe is more similar to each other than a triplet of embeddings for different food images, food ingredients, and food recipes.

10. The apparatus of claim 8 , wherein the processor analyzes food nutrients in the food image using at least one of the inferred food ingredient, the generated food recipe, and the generated food name.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (50)