Shot representation as support to scene segmentation

Author(s)

Tamires Tessarolli de Souza Barbieri

• Document type: Master's Dissertation
• Press: São Carlos.
• Institution: Universidade de São Paulo (USP). Instituto de Ciências Matemáticas e de Computação (ICMC/SB)
• Defense date: 2014-12-04

Examining board members

Rudinei Goularte; Celso Alberto Saibel Santos; Ricardo da Silva Torres

• Advisor: Rudinei Goularte

Abstract

The Content Personalization area has been the focus of recent researches in Computer Science and the automatic scene segmentation of digital videos is an important field supporting the composition of personalization services, such as content recommendation or summarization. One of the main approaches for scene segmentation is based on the clustering of related shots. Thus, in order to this process to be successful, is necessary to properly represent shots. However, we can see that the works reported on the literature have left this topic in backgroud. Therefore, this work aims to develop a method based on frames visual features, which enables to improve video shots representation and, consequently, the performance of scene segmentation techniques. (AU)

• FAPESP's process: 12/19025-0 - Shot Descriptors Supporting Scene Segmentation
• Grantee: Tamires Tessarolli de Souza Barbieri
• Support Opportunities: Scholarships in Brazil - Master