A Non-template Based Automatic Landmarking on 3D Face Data
SukTing Pui, Jacey-Lynn Minoi · 2019
The standard initial stage for the extraction of information from human face image data is the detection of key anatomical landmarks, which is a vital stage for several face recognition, facial analysis and synthesis applications. Locating facial landmarks in images is an important task in image processing, and detecting these landmarks automatically still remains a challenge. The appearance of facial landmarks may vary tremendously due to facial expressions and pose variations. Detecting and localising landmarks from raw face data are often performed manually by trained and experienced scientists or clinicians, and this process is laborious and tedious. In order to overcome these challenges, a novel non-template based automatic landmarking method on 3D face data is presented. The geometric approach is employed through utilising the surface curvatures, primitive surfaces information to detect and extract potential features. Subsequently, K-means clustering is applied to categorise and obtain the centroid of extracted features, which is later use to estimate and localise facial landmarks. The overall performance and accuracy of the proposed approach demonstrates the effectiveness and robustness of its method. Results of 95.6% of accurately locating the feature proved the effectiveness of the algorithm.