|
V7 N2 Paper 1
|
Annals of the MS in Computer Science and Information Systems at
UNC Wilmington
|
Fall 2013
|
An Open Source Extensible Automatic Landmarking System using Dynamic Active Shape Models (DASM)
David Macurak
Committee
Abstract
This thesis proposes a modern implementation of an application designed to perform automatic landmarking based on Active Shape Models. The software is designed to locate and precisely identify the features of an object, based on its known structure. In this particular study, the process of extracting features from frontal facial images is explored. The goal of which is to achieve accurate landmarks which can be applied to further processing techniques such as facial, expression, or gender recognition, etc. All of these techniques are reliant upon accurate annotations of features in order to produce meaningful results. Manually locating these landmarks can be a tedious and slow process, which is why automatic techniques such as the one discussed in this paper have been developed. Based on a comparison of similar automatic techniques, the open source implementation published in this work shows promising results, with many facets for future improvements.
download
(pdf)
Recommended Citation:
Macurak, D., Ricanek, K, Cummings, J., Simmonds, D. (2013) An Open Source Extensible Automatic Landmarking System using Dynamic Active Shape Models (DASM). Annals of the Master of Science in Computer Science and Information Systems at UNC Wilmington, 7(2) paper 1. http://csbapp.uncw.edu/data/mscsis/full.aspx.
|
V7 N2 Paper 1
|
Annals of the MS in Computer Science and Information Systems at
UNC Wilmington
|
Fall 2013
|
