This project is supported by the National Fund for Innovation and Scientific and Technological Development (FONDOCyT for its acronym in Spanish) from the Ministry of Higher Education, Science and Technology of the Dominican Republic (FONDOCYT 2018-19-3A1-107). The project is in collaboration with the Universidad Autonoma de Santo Domingo (UASD).
Project GitHub Repository HERE
The slides that provide an overview of the project and its objectives can be found HERE
You can watch a video of the implementation of an object detection algorithm on video footage of accidents from the Dominican Republic (HERE, viewer discretion is advised)
ABSTRACT
The objective of this project is to develop a method based on Machine Learning algorithms to detect traffic accidents in real-time with the use of traffic cameras. Early accident detection systems can save lives because, after an accident, a quick medical response can mean the difference between life and death. In addition, these systems could mitigate the economic impact of accidents by reducing bottlenecks and possible secondary incidents. Although the existing 9-1-1 System allows detecting accidents and providing medical assistance faster than ever before, accident detection technologies could help the system improve its response time and reduce its costs. Unfortunately, existing accident detection methods tend to have low accuracy and high levels of false alarm in complex urban environments, such as those prevailing in the Dominican Republic. Given these limitations and the impact that accident detection systems will have on the Dominican society, this project aims to develop a hybrid multistage method based on Machine Learning and Deep Neural Network algorithms. This method will help detect accidents in real-time, helping save lives, reduce bottlenecks, and reduce the economic impact of accidents. In addition, this proposal will help develop talent in the area of Artificial Intelligence, which is scarce in the Dominican Republic, despite the fact that it has been postulated as an essential area for technological innovation.
