| Grant number: | 20/06950-4 |
| Support Opportunities: | Research Grants - Problem-Oriented Research Centers in São Paulo |
| Start date: | February 01, 2024 |
| End date: | January 31, 2029 |
| Field of knowledge: | Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques |
| Principal Investigator: | João Eduardo Ferreira |
| Grantee: | João Eduardo Ferreira |
| Host Institution: | Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil |
| City of the host institution: | São Paulo |
| Principal investigators | Leliane Nunes de Barros ; Roberto Hirata Junior |
| Associated researchers: | Calton Pu ; Denis Deratani Mauá ; José Antonio Visintin |
| Associated scholarship(s): | 25/11717-0 - Phase 1: Intelligent Real-Time Entity and Incident Recognition by Deep Learning,
BP.TT 25/11421-4 - Automated Intelligent Monitoring System (AIMS) - Phase 2: Intelligent Real-Time Entity and Incident Recognition by Deep Learning, BP.TT 25/11422-0 - EBKA Approach to Live Information Monitoring, BP.TT |
Abstract
The GDP of the state of São Paulo reached R$ 2.38 trillion in 2019, with a 2.5% increase compared to 2018. Associated with this growth, there is a drop in some crime indicators, supported by the use of a semi-automated image monitoring system. However, there are still significant demands for the intelligent automation of the live video monitoring process in the areas of public security, intelligent transport and epidemiological surveillance. With hundreds of millions of video cameras in several cities around the world, human operators remain the main users of video data whose volume is growing explosively. This situation must change so that automated monitoring of live video in open spaces can become a cornerstone of operational planning for public security, intelligent transport and epidemiological surveillance. Despite the success in closed environments, such as airports and shopping malls (for example, in face recognition), the automated recognition of events in open spaces remains a major, important and urgent research and technological development challenge. The technical and fundamental difficulty arises in view of the occurrence of unprecedented events in open environments, exceeding the knowledge of the Machine Learning (ML) classifiers, trained and tested in fixed data sets, such as CIFAR (Canadian Institute for Advanced Research). Specifically, new entities and events, which appear as outliers, are generally considered noise by these classifiers. A tragic example of this limitation was the accident caused by Uber's automatic driving system (Arizona, 2018), when a pedestrian walking with a bicycle was fatally run over. The press described the cause of the accident as: "Uber software does not recognize human beings outside pedestrian crossings". This proposal describes the project called Automated Intelligent Monitoring System (AIMS), which aims to develop research, effective technologies and human resources training for real-time monitoring of live video, based on the premise of Evidence Based Knowledge Acquisition (EBKA). With the partnership that includes the University of São Paulo (USP), the Military Police of the State of São Paulo (PMESP), the companies Avantia and Equinix and the Eldorado Institute, the AIMS project will contribute to the detection and tracking of objects, such as people and vehicles, and analysis of events of interest in order to provide subsidies for efficient operational planning of public security, intelligent transport and also epidemiological surveillance. (AU)
| Articles published in Agência FAPESP Newsletter about the research grant: |
| More itemsLess items |
| TITULO |
| Articles published in other media outlets ( ): |
| More itemsLess items |
| VEICULO: TITULO (DATA) |
| VEICULO: TITULO (DATA) |