Publicação
Mobile application to identify recyclable materials
| Resumo: | This dissertation proposes a system to help the consumer recycle efficiently. The system is composed by a mobile application that can capture images of waste and classify their category through the usage of a machine learning model. Furthermore, this application can communicate with a server to update the model with new improved versions and also upload the images to the server in order to contribute to the creation of more precise model versions. The system has been validated by a fully working prototype. Although the proof of concept has been achieved, with some types of waste items correctly categorized, the machine learning model produced is not precise enough to be used in real-life scenarios, that is, for any type of waste. The main contributions of this study are a compendium of information in the area of computer vision and machine learning to categorize waste, and a working prototype system that utilizes crowdsourcing and machine learning elements to help the consumer recycle more efficiently. |
|---|---|
| Autores principais: | Sequeira, António Francisco Serol |
| Assunto: | Computer vision Machine learning Recycling Crowdsourcing Waste management Visão computacional Aprendizagem automática Reciclagem Contribuição colaborativa Gestão de resíduos |
| Ano: | 2020 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | ISCTE |
| Idioma: | inglês |
| Origem: | Repositório ISCTE |
| Resumo: | This dissertation proposes a system to help the consumer recycle efficiently. The system is composed by a mobile application that can capture images of waste and classify their category through the usage of a machine learning model. Furthermore, this application can communicate with a server to update the model with new improved versions and also upload the images to the server in order to contribute to the creation of more precise model versions. The system has been validated by a fully working prototype. Although the proof of concept has been achieved, with some types of waste items correctly categorized, the machine learning model produced is not precise enough to be used in real-life scenarios, that is, for any type of waste. The main contributions of this study are a compendium of information in the area of computer vision and machine learning to categorize waste, and a working prototype system that utilizes crowdsourcing and machine learning elements to help the consumer recycle more efficiently. |
|---|