Germany | Bayern

Zurück zur Suche

Internationale Partnersuche

Innovation & Technologie Angebot

French Technology Transfer Office proposes a machine learning system that identifies food items from any dish picture

Country of Origin: France
Reference Number: TOFR20180306001
Publication Date: 7 March 2018

Summary

The French TTO (Technology Transfer Office) is acting on behalf of an established public laboratory of the Paris region that has developed a machine learning based application that can analyse submitted images of food dishes for recognition.
The French public research centre is looking for partners for a technical cooperation or a research cooperation agreement.

Description

The French TTO (Technology Transfer Office) is acting on behalf of an established public laboratory of a Paris Region university.

A French academic laboratory has developed new methods for semantic image annotation. This topic is extensively studied for more than a decade now due to its large number of applications in areas as diverse as Information Retrieval, Computer Vision, Image Processing, and Artificial Intelligence. The recognition of object categories is one of the most challenging problem in the computer vision field, particularly in food image recognition, one of the promising applications.

*Innovative solution

The solution is an application (currently a web search engine) designed to retrieve, filter and classify images of dish recipes. It can run on a conventional mobile phone. It can recognize food items (from a photo of the meal taken by the user) and suggests the most relevant recipes based on a huge machine-learning dataset.

The technologies involved are :
• Eye-tracking interactive learning (data annotation & selection of relevant areas for image recognition),
• Deep-learning and new bio-inspired representations (biologically inspired networks that attempt at mimicking the primal areas of the virtual cortex),
• Web filtering for food annotation.

* Market Challenges :

Cooking and food are connected to many essential links of human life like wellness, health, products and bio-environment and ecology. Many websites have developed specific search engines and services for cooking recipes, some with crowd participation.

With the transformation of the digital society, the market is mature enough for hight technology image analysis, especially for retrieving images illustrating recipes. Food category classification is also a key challenge.

* Suggested applications :

• Recipe recommendation,
• Service to gourmet restaurants (picture of the gourmet dish and suggestion of the chef's recipe),
• Analysis of consumer trends in the Food-tech field,
• Quick payment in cafeteria by image analysis of trays,
• etc.

* Partnership :

The French TTO is able to provide technical and legal assistance to facilitate the eventual partnership. In case of business potential, the prototype can be partially funded by the TTO.

The partner sought could be a university or a company interested in :
- a research cooperation agreement on the actual software for some applications, or
- a technical cooperation agreement if the industrial application can benefit directly from the actual application.

*keywords :

#image processing
#computer vision
#deep learning
#pattern recognition
#machine learning
#computational cooking

Advantages and Innovations

Innovation :
• the duration of one image analysis is less than 2.5 seconds,
• more than 60% of positive results,
• need only 15 fixation points per image for recognition.


Advantages :
• Huge dataset and powerful image representations (101 food categories, each of them constituted by 800 to 950 images),
• Many potential applications and business models,
• Powerful classification algorithms,
• Expertise of the team in image processing, computer vision and deep learning.

Stage Of Development

Prototype available for demonstration

Stage Of Development Comment

A functional prototype has been developed and can be tested online.

Requested partner

The partner sought could be a food-tech company interested in a technological cooperation agreement, or a research cooperation agreement if any functionnalities have to be added. The eventual partner should have a clear industrial application in mind, and should have a strong technical background to co-develop the software with the laboratory as the prototype needs improvement.

Kooperationsanfrage stellen