Scribd
Upload
29 views14 pages

Food Recipe Generator

The document outlines a project aimed at developing a system that generates recipes from food images using convolutional neural networks (CNNs) and transformers. It highlights the limitations of existing systems that primarily rely on text or voice input, and proposes a solution that allows users to upload food images to receive ingredient lists and cooking instructions. The project aims to simplify cooking for users with varying levels of culinary knowledge and includes plans for future enhancements such as support for regional foods and multi-language options.

Uploaded by

sasankreddy.slvr
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
29 views14 pages

Food Recipe Generator

The document outlines a project aimed at developing a system that generates recipes from food images using convolutional neural networks (CNNs) and transformers. It highlights the limitations of existing systems that primarily rely on text or voice input, and proposes a solution that allows users to upload food images to receive ingredient lists and cooking instructions. The project aims to simplify cooking for users with varying levels of culinary knowledge and includes plans for future enhancements such as support for regional foods and multi-language options.

Uploaded by

sasankreddy.slvr
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

BAPATLA ENGINEERING COLLEGE :: BAPATLA

SECOND SHIFT POLYTECHNIC


Department Of Computer Engineering

RECIPE GENERATION FROM


FOOD IMAGE
Under The Guidance Of : Presented By :

E . Divya Kanthi M . Rupa : 23273-CM-049


Head Of The Department P . Sai Chand : 23273-CM-063
(Assistant Professor) U . Ravi Chandra : 23273-CM-080
A . Revanth : 23273-CM-007
CONTENTS
 Abstract

 Objective Of The Project

 Existing System

 Drawbacks of the Existing System

 Literature Review

 Proposed System

 Front-end Design

 Conclusion
Abstract
1. The system is designed to automatically generate
recipes from images of food.

2. It takes a food image as input and produces the


ingredients and step-by- step cooking instructions.

3. It uses convolutional neural networks (CNNs) to


understand the image and Transform to generate the
recipe text.

4. This system can be useful for home cooks, diet


monitoring, and food identification.
Objective Of The Project
1. Develop a system to generate a full recipe from a food image.

2. Identify ingredients in the dish using image analysis.

3. Generate easy-to-follow cooking steps based on the dish.

4. Help users recognize and cook food without basic knowledge.

5. Simplify cooking and support diet tracking using AI.


E xisting System
1. Most recipe recommendation systems depend on text or
voice search, not images.

2. Some apps suggest recipes from typed ingredients, but not


visual input.

3. Few systems, like Pic2Recipe, explore image-based recipe


generation.

4. Real-world use and accuracy of such systems are still


limited.
Drawbacks Of the Existing System
1. Rely on text or voice input, not visual input.

2. Food recognition from images is often inaccurate.

3. No proper link between image analysis and recipe creation.

4. Few datasets connect images with full recipes.

5. Existing models (e.g.,Pic2Recipe) show low real-world


accuracy.
Literature Review
1. Inverse Cooking: uses transformers to predict ingredients and generate
recipes directly from food images.

2 . FIRE: applies an end-to-end transformer model to jointly predict


ingredients and cooking instructions.

3 . AdaMine: uses adaptive triplet mining to learn strong cross-modal


embedding for image-to-recipe retrieval.

4 . Vision-Language Cross-Modal Transformers: combine Vision


Transformers and BERT for advanced image-text learning.

5 . GPT-4V: is a multimodal version of GPT-4 that processes images and


generates fluent recipes using vision-language understanding.
Proposed System
1. The user uploads a food image.

2. A CNN model extracts important visual features from the


image.

3. A model predicts ingredients, and a Transformer generates the


recipe.

4. The system shows ingredients and step-by-step cooking


instructions.
Front-end Design
1. The user opens the app and sees the upload interface.

2. The user uploads a food image.

3. The image is sent to the backend.

4. The frontend waits for the response.

5. Predicted ingredients and recipe are displayed.

6. The user sees results with proper feedback (like loading or


error messages).
Conclusion
1.The system generates recipes from food images using
CNNs and transformers.

2. It predicts ingredients and provides step-by-step


cooking instructions.

3. Useful for home cooks, bloggers, and health-conscious


users.

4. Future improvements: regional foods, diet options,


voice/camera input, and multi-language support.

You might also like