CrowdPet: applying computer vision in the identification of lost pets

• Grant number: 16/01957-5
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: February 01, 2017
• End date: October 31, 2017
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Fabio Rogério Piva
• Grantee: Fabio Rogério Piva
• Company: SciPet Soluções em Inovação Tecnológica Ltda. - ME
• City: Campinas
• Pesquisadores principais: Allysson Flavio Melo de Oliveira ; Fernanda Alcântara Andaló ; Orlando Volpato Filho
• Associated researchers: Eduardo Alves Do Valle Junior ; Mauricio Lisboa Perez ; Rafael de Oliveira Werneck
• Associated research grant(s): 17/16244-7 - CrowdPet: automatic identification of pets in images, AP.PIPE
• Associated scholarship(s): 17/01431-6 - CrowdPet: applying computer vision in the identification of lost pets, BP.PIPE

Abstract

Brazil, just like several other developing countries, has a significant problem regarding street animals population. Aside the cruel conditions, this situation represents risks for public health, since the animals may become sick and procreate at alarming rates. We intend to employ technological solutions -- specifically, computer vision and machine learning methods -- in order to approach this urban problem in a collaborative fashion, making it easier for citizens to take part on the process of identification and mapping of those animals. This project is proposed in the context of the CrowdPet service, which will be made available to users in the form of a smartphone App, thus allowing for the easy determination of lost pets. On one hand, pet owners (aka, tutors) can submit pictures of their lost pets to the system, while regular users (aka, volunteers) will submit pictures they take from street animals they encounter on their daily commute. Our main goal is to contribute with computational methods for the visual recognition of animals, so that the process of matching lost and found animals can be made in an automatic, scalable and efficient fashion. (AU)

Filed patent(s) as a result of this research project