Please use this identifier to cite or link to this item:
http://hdl.handle.net/2445/148059
Title: | Detecció de safates en temps real per sistema d’autoservei en restaurants |
Author: | Vancells Lujan |
Director/Tutor: | Bolaños Solà, Marc |
Keywords: | Visió per ordinador Reconeixement de formes (Informàtica) Programari Treballs de fi de grau Aprenentatge automàtic Xarxes neuronals (Informàtica) Gestió de restaurants Aliments Food Computer vision Pattern recognition systems Computer software Machine learning Neural networks (Computer science) Bachelor's theses Restaurant management |
Issue Date: | 27-Jun-2019 |
Abstract: | [en] Daily, a lot of people use self-service restaurants and fast food franchises. These services feed thousands of people both in the business sector and the hospitality industry. Lately, public awareness of the importance of good nutritional habits has been significantly increased, mostly thanks to the growth of social networks. These trends offer a great opportunity to an innovative technology, the recognition of food images. Computer vision which allows us to obtain quantitative and qualitative information from images, could benefit both businesses and the consumer, providing a positive impact in the sector. With the great advances on deep learning techniques of the last years, we can classify and identify objects in any type of images, using computer vision. Applying this to the specific case of self-service restaurants, it would allow to optimize the management of the foods used in this type of business, as well as the automation of the payment methodologies of these services. In this thesis we propose a method to facilitate the processing of images containing trays with food, creating a model capable of identifying whether there is a tray in an image or not. To do this, we have created a collection of images to create such a model, using convolutional neural networks. This model will be applied in a real case, specifically in a food recognition system at a commercial level. The proposed model improves the results obtained by the previous technology by 38%. Enhancing the F1 Score from 0.529 to 0.911. |
Note: | Treballs Finals de Grau d'Enginyeria Informàtica, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2019, Director: Marc Bolaños Solà |
URI: | http://hdl.handle.net/2445/148059 |
Appears in Collections: | Treballs Finals de Grau (TFG) - Enginyeria Informàtica |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
memoria.pdf | Memòria | 30.28 MB | Adobe PDF | View/Open |
This item is licensed under a Creative Commons License