Automatic 3D landmarking model using patch‐based deep neural networks for CT image of oral and maxillofacial surgery

Abstract

Background: Manual landmarking is a time consuming and highly professional work.Although some algorithm-based landmarking methods have been proposed, they lack flexibility and may be susceptible to data diversity. Methods: The CT images from 66 patients who underwent oral and maxillofacial surgery (OMS) were landmarked manually in MIMICS. Then the CT slices were exported as images for recreating the 3D volume. The coordinate data of landmarks were further processed in Matlab using a principal component analysis (PCA) method. A patch-based deep neural network model with a three-layer convolutional neural network (CNN) was trained to obtain landmarks from CT images. Results: The evaluating experiment showed that this CNN model could automatically finish landmarking in an average processing time of 37.871 seconds with an average accuracy of 5.785 mm. Conclusion: This study shows a promising potential to relieve the workload of the surgeon and reduces the dependence on human experience for OMS landmarking. K E Y W O R D S 3D cephalometry, automatic landmarking, convolutional neural network, machine learning, oral and maxillofacial surgery

Applications of artificial intelligence and machine learning in orthodontics: a scoping review

“…Selection of orthodontic appliance- type of headgear 1 [ 63 ] p. Quantification of sagittal skeletal discrepancy 1 [ 49 ] q. For cases suitable for fixed mechanotherapy 1 [ 55 ] r. Selection of patients suitable to be treated with removable orthodontic appliances 1 [ 56 ] s. Class II division 1 malocclusion 1 [ 57 ] t. Broad-based 1 [ 79 ] Automated cephalometric landmarking and/or analysis and/or classification a. Lateral cephalogram 12 [ 27 , 32 , 35 , 37 , 64 , 65 , 69 , 70 , 74 , 75 , 77 , 78 ] b. CBCT images 6 [ 31 , 58 – 61 , 76 ] c. Frontal cephalogram 1 [ 19 ] Assessment of growth and development a. Cervical vertebra maturation 1 [ 18 ] b. Broad-based 3 [ 30 , 39 , 73 ] Evaluation of treatment outcome- orthognathic surgery on facial appearance/ attractiveness and/or age perception 2 [ 36 , 38 ] Miscellaneous …”

“…The use of CBCT for cephalometric analysis has now become commonplace. Various studies [31,[58][59][60][61]76] that have employed AI and ML techniques for automatic landmark detection and analysis have shown that the results obtained are as accurate and less timeconsuming as compared to those obtained with manual analysis. At least one study [19] included in this review compared frontal cephalometric landmarking ability of humans versus that of artificial neural networks and the results showed that ANNs could achieve accuracy comparable to humans in placing cephalometric points, and in some cases surpasses the accuracy of inexperienced doctors (students, residents, graduate students).…”

Robotic Applications in Orthodontics: Changing the Face of Contemporary Clinical Care

“…Robots for dental articulation can eradicate technical difficulties of duplication positions and motions encountered in the classical articulator, thus saving time and producing more precise occlusal relationships [17]. Robots for automated cephalometric landmarking save time and decrease the dependence on professional experience [62,63]. Titration of oral appliances for treatment of OSA patients in a single night within minutes was introduced by robots to overcome the time-consuming trial-and-error procedures.…”

“…A new approach for automatic 3D cephalometric annotation using shadowed 2D image-based machine learning was proposed to overcome the existing serious difficulties in handling high-dimensional 3D CT data, achieving an average point-to-point error of 1.5 mm for seven major landmarks [62]. Also, another study using a patch-based iterative network with a three-layer CNN architecture for automatic landmarking of a CT image showed that landmarks can be automatically calculated in 37.871 seconds with an average acceptable accuracy of 5.785 mm [63].…”

Machine learning in dental, oral and craniofacial imaging: a review of recent progress

“…Some structures like gonion, porion and others seem to be points with imperfect accuracy. In addition to algorithm insufficiency and manual errors, inexact anatomical positions and complex definitions are possible causes of this loss of accuracy ( Ma et al, 2020 ; Montúfar, Romero & Scougall-Vilchis, 2018 ). However, only a few studies have been reported in the three-dimensional field of imaging, which suggests that it is still at the initial stages.…”