Image segmentation and feature extraction by regions for the creation of multi-instance learning scenarios

• Grant number: 14/14557-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: August 01, 2015
• End date: November 30, 2016
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Rosane Minghim
• Grantee: Leo Sampaio Ferraz Ribeiro
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated research grant: 11/22749-8 - Challenges in exploratory visualization of multidimensional data: paradigms, scalability and applications, AP.TEM
• Associated scholarship(s): 15/26050-0 - Multiple-instance image ranking for sketch-based image retrieval, BE.EP.IC

Abstract

There are image classification problems in which each image are represented by regions of interest; for each region a series of features are extracted. As a result a set of feature vectors are available, and it is necessary to assign a label to this set of instances. The Multi-Instance Learning studies the problem in which each object is describes as a bag (set of instances). This project aims the study of color spaces and segmentation methods that allow the extraction of feature vectors that can represent the images successfully. We are going to investigate the color spaces RGB, Lab and HSV, and also color, texture and shape features. We expect to create datasets to multi-instance learning, that can be visualized using methods of multiscale visualization, helping the study of instance and bag spaces.