Visual samples manipulation to image object recognition

• Grant number: 12/12334-8
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: August 01, 2012
• End date: July 31, 2013
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Moacir Antonelli Ponti
• Grantee: Rodrigo Vicente Casagrande Beber
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

Object recognition in scenes is a topic of recent interest due to the technologies and data sets currently available for image and video data. However, in real application data there is often a significant variation within the samples available for each object (class) to be recognized. This issue can hamper the use of regular supervised learning methods. This project aims to study algorithms to balance the image sets using automatic sample generation and removal of redundant samples, so that the imbalance is compensated. It is expected that this method allow the use of more meaningful samples and improve the accuracy of the object recognition system. (AU)