Cephalometric Landmarks3/28/2021
For an overall evaluation of the clinical viability of automated landmarking of this extended method, in this investigation the following null hypothesis was tested: there is no statistically significant difference in accuracy between the 10 landmarks automatically located by this approach and the true location of every landmark. 2. Materials and Methods 2.1. Image Sample Forty-one lateral cephalometric radiograph files taken at the Orthodontic Department of Policlinico, University Hospital of Catania, Italy, were used in this study.Sofia 78, 95123 Catania, Italy 2 Dipartimento di Ingegneria Informatica e Telecomunicazioni, Universit di Catania, Viale A.
Doria 6, 95125 Catania, Italy Show more Academic Editor: Rita Casadio Received 16 Feb 2009 Revised 16 May 2009 Accepted 18 Jun 2009 Published 10 Sep 2009 Abstract Several efforts have been made to completely automate cephalometric analysis by automatic landmark search. Cephalometric Landmarks Manual Identification InHowever, accuracy obtained was worse than manual identification in every study. The analogue-to-digital conversion of X-ray has been claimed to be the main problem. Therefore the aim of this investigation was to evaluate the accuracy of the Cellular Neural Networks approach for automatic location of cephalometric landmarks on softcopy of direct digital cephalometric X-rays. Forty-one, direct-digital lateral cephalometric radiographs were obtained by a Siemens Orthophos DS Ceph and were used in this study and 10 landmarks (N, A Point, Ba, Po, Pt, B Point, Pg, PM, UIE, LIE) were the object of automatic landmark identification. The mean errors and standard deviations from the best estimate of cephalometric points were calculated for each landmark. Differences in the mean errors of automatic and manual landmarking were compared with a 1-way analysis of variance. The analyses indicated that the differences were very small, and they were found at most within 0.59 mm. Furthermore, only few of these differences were statistically significant, but differences were so small to be in most instances clinically meaningless. Investigations on softcopy of digital cephalometric X-rays, to search more landmarks in order to enable a complete automatic cephalometric analysis, are strongly encouraged. Introduction Since the introduction of the cephalometer in 1931 1, cephalometric analysis has become an important clinical tool in diagnosis, treatment planning, evaluation of growth, or treatment results and research 2, 3. In this respect, automatic cephalometric analysis is one of the main goals, to be reached in orthodontics in the near future. Accordingly, several efforts have been made to automate cephalometric analysis 4. The main problem, in automated cephalometric analysis, is landmark detection, given that the measurement process has already been automated successfully. Different approaches that involved computer vision and artificial intelligence techniques have been used to detect cephalometric landmarks 5 22, but in any case accuracy was the same or worse than the one of manual identification; for a review see Leonardi et al. None of the proposed approaches solves the problem completely, that is, locating all the landmarks requested by a complete cephalometric analysis and with accuracy suitable to clinical practice. It should be emphasized that reliability in the detection of landmarks is mandatory for any automatic approach, in order to be employed for any clinical use. As previously stated 4, among the possible factors that reduce reliability the loss of image quality, inherent to digital image conversion and compression in comparison with the original radiograph, has been claimed 3, 23, 24. In fact, this analogue-to-digital conversion (ADC) results in the loss and alteration of information due to the partial volume averaging; consequently many edges are lost or distorted. To the best of our knowledge, every study on automatic landmarking has been carried out on scanned lateral cephalograms transformed into digital images 4, and this could explain in some way the inaccuracies of automatic location compared to the manual identification of landmarks. Recently, a new hybrid approach, which is based on Cellular Neural Networks (CNNs), has been proposed for automatic detection of some landmarks 21, 22. Results of evaluation of the methods performance on scanned cephalograms were promising; nevertheless, for some landmarks the error in the location was often greater than the one of manual location. Due to the promising results already obtained with CNNs 21, 22, the aim of this study was to evaluate the accuracy of the CCNs-based approach for the automatic location of cephalometric landmarks on direct digital cephalometric X-rays. Thus the method proposed in 21, 22 has been extended in two respects: by improving the algorithms employed to detect 7 landmarks and by developing the algorithms needed to locate 3 additional landmarks (Porion, Basion, and Pterygoid point); of these latter, two especially difficult landmarks (Basion and Pterygoid point) that are used in the most common cephalometric analysis were located for the first time in literature. For an overall evaluation of the clinical viability of automated landmarking of this extended method, in this investigation the following null hypothesis was tested: there is no statistically significant difference in accuracy between the 10 landmarks automatically located by this approach and the true location of every landmark. Materials and Methods 2.1. Image Sample Forty-one lateral cephalometric radiograph files taken at the Orthodontic Department of Policlinico, University Hospital of Catania, Italy, were used in this study.
0 Comments
Leave a Reply.AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |