by Philip Limbeck
Abstract:
The human face provides a rich source of information which can be exploited to diagnose facial impairments. Facial palsy is one of these impairments, and is caused by restrictions of the nerval actuation of muscles responsible for facial expressions. The main symptoms of this condition are asymmetrical facial movement and partial facial paralysis. To measure its progress, physicians require clinical measures extracted from those locations of the face which provide most information about the facial expression. Small artificial markers indicate these locations by being placed on the patient's face before an evaluation session. A video of the patient is recorded which is used to localize these markers in every frame. This task is currently performed manually by an operator and can take up to 5 hours for a single video. Object tracking refers to the estimation of the position of objects from an image sequence. Illumination and occlusion are considered as the main problems when tracking artificial objects. Natural objects, such as the human face, have a high potential for deformation and are characterized by an irregular texture. As not only one, but multiple markers have to be tracked simultaneously, additional diffculty is imposed by ensuring that markers can be uniquely identified. The thesis explores the possibility of tracking these markers semi-automatically by a applying a sequential Bayes estimation technique, which assesses a set of hypothesis using their congruence with the target model. Hence, the location of each marker can be accurately estimated and occlusions handled efficiently. To improve the accuracy and to reset lost markers, the clinical operator can interact with the tracking system. The results showed that our chosen methods are superior when compared with traditional trackers which use only a single hypothesis concerning the marker locations, while at the same time being able to preserve an accuracy comparable to manual tracking.
Reference:
Interactive Tracking of Markers for Facial Palsy Analysis (Philip Limbeck), Technical report, PRIP, TU Wien, 2012.
Bibtex Entry:
@TechReport{TR128,
author = "Philip Limbeck",
title = "Interactive Tracking of Markers for Facial Palsy Analysis",
institution = "PRIP, TU Wien",
number = "PRIP-TR-128",
year = "2012",
url = "https://www.prip.tuwien.ac.at/pripfiles/trs/tr128.pdf",
abstract = "The human face provides a rich source of information which can be exploited to diagnose facial
impairments. Facial palsy is one of these impairments, and is caused by restrictions of the nerval
actuation of muscles responsible for facial expressions. The main symptoms of this condition are
asymmetrical facial movement and partial facial paralysis. To measure its progress, physicians
require clinical measures extracted from those locations of the face which provide most information
about the facial expression. Small artificial markers indicate these locations by being placed on
the patient's face before an evaluation session. A video of the patient is recorded which is used to
localize these markers in every frame. This task is currently performed manually by an operator
and can take up to 5 hours for a single video. Object tracking refers to the estimation of the
position of objects from an image sequence. Illumination and occlusion are considered as the
main problems when tracking artificial objects. Natural objects, such as the human face, have a
high potential for deformation and are characterized by an irregular texture. As not only one, but
multiple markers have to be tracked simultaneously, additional diffculty is imposed by ensuring
that markers can be uniquely identified. The thesis explores the possibility of tracking these
markers semi-automatically by a applying a sequential Bayes estimation technique, which assesses
a set of hypothesis using their congruence with the target model. Hence, the location of each
marker can be accurately estimated and occlusions handled efficiently. To improve the accuracy
and to reset lost markers, the clinical operator can interact with the tracking system. The results
showed that our chosen methods are superior when compared with traditional trackers which use
only a single hypothesis concerning the marker locations, while at the same time being able to
preserve an accuracy comparable to manual tracking.",
}