• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.9
• /
• pp.3685-3700
• /
• 2015

To avoid influences caused by pose, illumination and facial expression variations, we propose a robust three-step algorithm based on ASM and POEM for facial landmark localization. Firstly, Model Selection Factor is utilized to achieve a pose-free initialized shape. Then, we use the global shape model of ASM to describe the whole face and the texture model POEM to adjust the position of each landmark. Thirdly, a second localization is presented to discriminatively refine the subtle shape variation for some organs and contours. Experiments are conducted in four main face datasets, and the results demonstrate that the proposed method accurately localizes facial landmarks and outperforms other state-of-the-art methods.

• Korean Journal of Head & Neck Oncology
• /
• v.22 no.1
• /
• pp.29-32
• /
• 2006

Background: The location of parotid gland tumors can influence the duration and the difficulty of the operation. If the information about tumor location was available preoperatively, it would allow accurate operative planning and counseling of patients in terms of the length of the operation and the potential morbidity. Methods: This study was based on a retrospective review of 100 patients with parotid gland tumors underwent parotidectomy from January 2000 to October 2005 at Yong-Dong Severance Hospital. Based on computed tomographic(CT) scan findings, 4 landmarks such as facial nerve(FN) line, Utrecht(U) line, Conn's are(CA), and retromandibular vein (RV) were drawn on the scans in same plane. The location of tumors were determined by the landmarks and confirmed by the operative findings. The accuracy of each landmarks was evaluated. To find out the accuracies according to tumor size, the tumors were divided into 2 groups; less than 2 cm and larger than 2 cm in diameter. Results: U line was the most accurate(94%), sensitive(89.3%) and specific(97.7%) in predicting tumor location of the parotid gland. However, in small tumors less than 2cm, FN line (p=0.022) and RV criteria (p=0.028) were more reliable in accuracy. Conclusion: CA, FN line, U line, and RV are all useful landmarks in preoperative prediction for the location of parotid gland tumors. However, U line was the most accurate, but we must consider that proper landmark should be used in prediction according to the size of tumor because the accuracy of landmark may change.