Adressing methods of Domain Adaptation between synthetic and real images in object detection

• Grant number: 22/15998-6
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: February 01, 2023
• End date: May 07, 2024
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computational Mathematics
• Principal Investigator: André Carlos Ponce de Leon Ferreira de Carvalho
• Grantee: Davi Fileti
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

The project consists in studying domain adaptation techniques towards image object detection. This would allow, for instance, the use of synthetic images for training neural networks. Utilizing this data alongside with real images, time and money would be spared and the scope of researchers and enterprises able to deploy this kind of model would expand significantly since this models will be easily trained. Self driving cars, detection of failures in transmission towers by drones, vigilance, detection of hematomas in animals in a zoo and many other applications would directly benefit from these domain adaptation techniques.