Bilkent University
Department of Computer Engineering

---

S E M I N A R

 

Generic Contextual Object Recognition and Semantic Scene Classification

 

Demir Gokalp
MSc. Student
Computer Engineering
Bilkent University

Semantic scene classification is a new field which categorizes images into physical or semantic categories. Classifying images using components like regions, pixels and objects is a challenging work because of the ambiguity of the visual data. These components (pixels, regions and objects) in natural images have reasonable spatial and geometric interactions among them. To make a general inference about the content of an image, we need to divide it into meaningfully labeled regions. Using context and spatial relationships between these regions, we can make inference about the image from specific to general. Compared to common approach which is applying different classifiers for each object or semantic class, this research aims to model more generic approaches comprising main classes using logical and probabilistic relationships of components in natural images.

 

DATE: April 10, 2006, Monday@ 16:40
PLACE: EA 502