Empirical data final project / Yaniv Bar, March 2012


A conceptual implementation of the following paper

[1] TextonBoost: Joint Appearance, Shape and Context Modeling for Multi-Class Object Recognition and Segmentation

By J. Shotton, J. Winn, C. Rother, and A. Criminisi (2006)

The paper investigates the problem of achieving automatic detection, recognition and segmentation of object classes in photographs. Mainly, given an image the system should automatically partition it into semantically meaningful areas each labeled with a specific object class.

--> Project Report <--

The project is accompanied with a short power point presentation.--> Project Presentation <--

Data Set:

The data which is used in the paper can be accessed freely for research purposes and it is mainly demonstrated by a 23-object class database of photographs of real objects viewed under general lighting conditions, poses and viewpoints.
The data set can be downloaded via the following link: Microsoft Image Understanding Research Data(Research data sub-section, 2nd version of the data set)

Code:

A self contained compressed .rar file containing all project files (relevant data, code and classification attempts.

--> Project Code <--



References:

[1] TextonBoost: Joint Appearance, Shape and Context Modeling for Multi-Class Object Recognition and Segmentation
[2] TextonBoost for Image Understanding: Multi-Class Object Recognition and Segmentation by Jointly Modeling Texture, Layout, and Context