Thursday, January 26, 2006
Landmarks - manual or automatic
My supervisor wanted me to implement an automatic landmarking, however as I was reading and reviewing on landmarks, I found a paper on automatic landmarking which is based on stereo method onto 3D object. The thought of the idea was indeed brilliant, however the more I read, I realised that setting your landmarks manually seems to be a better option. This is to enable in creating an accurate model and any errors on correspondence points can be reduced. Not forgetting, in implementing landmark, the number of landmarks is also important.
The few questions to ponder and answer:
1. Study and compare the accuracy and the effectiveness of automatic landmarking and manual landmarking.
2. Study and identify the framework for landmarking faces.
3. Identify how many landmarks were used and justify the reasons to the choice.
4. Identify the step-by-step method in implementing landmarking and the next process.
Automatic landmarking
Author: A Naftel
Title: Visualisation of Shape Change by Stereo-assisted landmark detection: an application to medical imaging (2003)
Abstract:
Techniques for the 3-D analysis of facial shape changes resulting from growth or surgical treatment have recently been proposed. These approaches rely on manual intervention for placing anatomical landmarks on 3-D facial models which is time consuming and error prone. In this paper, an automated approach to landmark extraction is described which uses a stereo-assisted active shape model. A landmark-based statistical analysis of facial shape change is then carried out using procrustes registration, principal component analysis and thin plate spline warping. The proposed method is validated both statistically and visually by characterizing shape changes induced by mandibular repositioning in a cross-sample of 20 orthodontic patients. It is shown that the method is capable of distinguishing between changes in facial morphology due to mandibular repositioning and changes due to other factors such as growth and normal variation within the patient sample.
Authors: Nafterl, Mao, Trenouth
Title: Stereo-assisted landmark detection for the anaysis of 3D facial shape changes, technical report (2002)
The few questions to ponder and answer:
1. Study and compare the accuracy and the effectiveness of automatic landmarking and manual landmarking.
2. Study and identify the framework for landmarking faces.
3. Identify how many landmarks were used and justify the reasons to the choice.
4. Identify the step-by-step method in implementing landmarking and the next process.
Automatic landmarking
Author: A Naftel
Title: Visualisation of Shape Change by Stereo-assisted landmark detection: an application to medical imaging (2003)
Abstract:
Techniques for the 3-D analysis of facial shape changes resulting from growth or surgical treatment have recently been proposed. These approaches rely on manual intervention for placing anatomical landmarks on 3-D facial models which is time consuming and error prone. In this paper, an automated approach to landmark extraction is described which uses a stereo-assisted active shape model. A landmark-based statistical analysis of facial shape change is then carried out using procrustes registration, principal component analysis and thin plate spline warping. The proposed method is validated both statistically and visually by characterizing shape changes induced by mandibular repositioning in a cross-sample of 20 orthodontic patients. It is shown that the method is capable of distinguishing between changes in facial morphology due to mandibular repositioning and changes due to other factors such as growth and normal variation within the patient sample.
Authors: Nafterl, Mao, Trenouth
Title: Stereo-assisted landmark detection for the anaysis of 3D facial shape changes, technical report (2002)
# posted by Jacey Lynn Minoi : 12:02 pm
