WO2023246463A1 - 口腔全景片的分割方法 - Google Patents
口腔全景片的分割方法 Download PDFInfo
- Publication number
- WO2023246463A1 WO2023246463A1 PCT/CN2023/097792 CN2023097792W WO2023246463A1 WO 2023246463 A1 WO2023246463 A1 WO 2023246463A1 CN 2023097792 W CN2023097792 W CN 2023097792W WO 2023246463 A1 WO2023246463 A1 WO 2023246463A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- oral
- segmentation
- computer
- physiological structure
- physiological
- Prior art date
Links
- 230000011218 segmentation Effects 0.000 title claims abstract description 59
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000013528 artificial neural network Methods 0.000 claims abstract description 21
- 230000000717 retained effect Effects 0.000 claims abstract description 3
- 210000004086 maxillary sinus Anatomy 0.000 claims description 20
- 210000000988 bone and bone Anatomy 0.000 claims description 19
- 210000000276 neural tube Anatomy 0.000 claims description 13
- 230000000877 morphologic effect Effects 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 description 11
- 238000004590 computer program Methods 0.000 description 7
- 238000012549 training Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000003325 tomography Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 210000002050 maxilla Anatomy 0.000 description 1
- 230000003239 periodontal effect Effects 0.000 description 1
- 238000010882 preoperative diagnosis Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000002601 radiography Methods 0.000 description 1
- 210000004872 soft tissue Anatomy 0.000 description 1
- 210000004357 third molar Anatomy 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 210000001782 transverse sinus Anatomy 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
Definitions
- the present application generally relates to a segmentation method for oral panoramic films.
- Oral panoramic radiography also known as curved tomography, is a simple and fast technology that uses the principles of narrow slit and arc orbit tomography to obtain a total image of all teeth and surrounding tissues through one imaging.
- the oral panoramic radiograph Based on the oral panoramic radiograph, it can accurately measure the tooth inclination angle, grasp the condition of periodontal soft tissue, tooth roots and alveolar bone, and accurately measure the anatomical shape, etc., providing a panoramic image basis for preoperative diagnosis analysis and treatment plan design.
- One aspect of the present application provides a computer-implemented segmentation method of an oral panorama, which includes: obtaining an oral panorama; using a trained first deep neural network to segment physiological structures in the oral panorama, wherein , segment the physiological structure with left and right parts according to the same category; and for the segmentation result of each physiological structure with left and right parts, only retain the two largest connected domains, and based on the two largest The left and right relative positions of the connected domains are further divided into left and right parts of the physiological structure.
- the first deep neural network may be a U-Net network.
- the physiological structures include: the maxillary sinus, the condylar process, the alveolar bone, and the mandibular neural canal, wherein the maxillary sinus, the condylar process, and the mandibular neural canal have two parts: left and right.
- the computer-executed segmentation method of the oral panorama further includes: using a trained second deep neural network to crop out a target region image from the oral panorama, and utilizing the first The segmentation of the deep neural network is performed based on the target area image, wherein the target area includes pre-selected physiological structures.
- the second deep neural network may be yolov3.
- the computer-executed segmentation method of the oral panorama further includes: using the trained second deep neural network to detect the first physiological structure region and the second physiological structure in the oral panorama. area; and correcting the direction of the target area image based on the detected positional relationship between the first physiological structure area and the second physiological structure area, and the distribution of the first physiological structure area or the second physiological structure area, The segmentation using the first deep neural network is performed based on the corrected direction target area image.
- the first physiological structure region and the second physiological structure region are expressed as rectangular boxes.
- the first physiological structure is alveolar bone
- the second physiological structure is the maxillary sinus
- the correction of the direction is based on the length and width direction of the detected alveolar bone area and the It is performed based on the positional relationship between the alveolar bone area and the maxillary sinus area.
- the computer-executed segmentation method of the oral panorama further includes: performing a morphological closure operation on the segmentation result.
- the computer-executed segmentation method of the oral panorama further includes: extracting contour points of the segmentation result after morphological closed operation; and based on the extracted contour points, using a spline curve The new contour line is obtained by fitting as the final segmentation result.
- Figure 1 is a schematic flow chart of a computer-executed segmentation method for oral panoramic films in one embodiment of the present application
- Figure 2A is an example of an oral panoramic photograph showing an interface of a computer program for segmenting an oral panoramic photograph in an embodiment of the present application
- Figure 2B is an image of a target area cropped from the oral panorama shown in Figure 2A displayed on an interface of the computer program;
- FIG. 2C is a segmentation result of the target area image shown in FIG. 2B displayed on an interface of the computer program.
- One aspect of the present application provides a computer-implemented segmentation method for oral panoramic films.
- FIG. 1 is a schematic flow chart of a computer-executed segmentation method 100 for an oral panoramic image in one embodiment of the present application.
- Another aspect of the present application provides a computer system for segmenting oral panoramic photos, which includes a storage device and a processor.
- the storage device stores a computer program that will be executed when run by the processor.
- the segmentation method 100 of the oral panoramic photograph is also provided.
- the oral panoramic radiograph to be processed is obtained.
- the segmentation method 100 of the oral panorama of the present application can not only process the original oral panorama, but also the reproduced oral panorama, for example, the picture obtained by using a smart mobile phone to re-photograph the paper oral panorama.
- FIG. 2A shows an interface of a computer program for segmenting an oral panoramic photograph according to an embodiment of the present application, showing an example of an oral panoramic photograph.
- a target area image is cropped from the oral panorama.
- the oral panoramic picture may be obtained by original shooting, or may be obtained by secondary shooting (for example, using a smart mobile phone to take the actual oral panoramic picture or the oral panoramic picture displayed on a computer screen).
- the oral panorama obtained by secondary shooting may include some unnecessary content, such as the scenery surrounding the original oral panorama during the secondary shooting, which may affect subsequent processing to a certain extent.
- images of target areas containing pre-selected physiological structures can be cropped from the oral panorama, and segmentation can be performed based on the cropped target area images.
- the target area may be a rectangle.
- a trained deep neural network can be used to detect the target area in the oral panorama, and crop the target area image based on the detection results.
- the deep neural network is hereinafter referred to as the target detection network.
- the target detection network in addition to the target area, can also be used to detect the maxillary sinus area and alveolar bone area in the oral panorama, and correct the direction of the target area image based on the detection results.
- the specific operations are as follows.
- the detection results of the maxillary sinus area and alveolar bone area are only used for direction correction, the detection accuracy is not high.
- the detection results of the maxillary sinus area and the alveolar bone area can be expressed by a rectangular frame.
- the alveolar bone area is generally rectangular, if the longitudinal size of the detected alveolar bone area is larger than the transverse size, the image is rotated 90 degrees counterclockwise.
- the maxillary sinus area should be above the alveolar bone area, if the maxillary sinus area is below the alveolar bone area, rotate the picture 180 degrees counterclockwise.
- a combination of other physiological structures can also be used to correct the direction of the target area image, for example, a combination of the maxillary sinus area and the mandibular neural canal area. , or a combination of the condylar area and the mandibular neural canal area, etc.
- the target detection network may use a yolov3 network. It can be understood that in addition to the yolov3 network, the target detection network can also use any other applicable deep neural network, such as SSD and Faster RCNN network.
- methods such as flipping and rotating 90 degrees can be used to augment the training set to improve the robustness of the target detection network.
- FIG. 2B is an image of a target area cropped from the oral panorama shown in FIG. 2A displayed on an interface of the computer program.
- segmentation of relevant physiological structures based on target area images has advantages in network training and segmentation accuracy, segmentation directly based on the original oral panorama is also feasible.
- a trained deep neural network can be used to segment the physiological structure in the target area image.
- the deep neural network is hereinafter referred to as a segmentation network.
- the segmentation network may be a U-Net network. It can be understood that in addition to the U-Net network, the segmentation network can also adopt any other applicable deep neural network, for example, semantic segmentation networks such as DeepLab or FCN.
- the following physiological structures can be segmented from the target area image: alveolar bone, maxillary sinus, condyle, and mandibular neural canal.
- the maxillary sinus, condyle, and mandibular neural canal have two parts, the left and right parts.
- the following methods can be used to process their segmentation results.
- the connected domains of the segmentation map are calculated, and the two largest connected domains are retained.
- the abscissas of the centers of the two connected domains are calculated. Based on the calculation results, the maxillary sinus, condyle, and mandibular neural canal are further divided into the left maxillary sinus, left condylar process, left mandibular neural canal, and right Lateral sinus, right condyle, right mandibular neural canal.
- the segmentation results are post-processed.
- the segmentation results obtained above may have defects such as holes and insufficiently smooth contours, so they can be post-processed accordingly.
- a morphological closing operation can be performed on the above segmentation results to eliminate possible A hole that can exist.
- FIG. 2C is a segmentation result of the target area image shown in FIG. 2B displayed on an interface of the computer program.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Dental Tools And Instruments Or Auxiliary Dental Instruments (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
本申请的一方面提供了一种计算机执行的口腔全景片的分割方法,其包括:获取口腔全景片;利用经训练的第一深度神经网络对所述口腔全景片中的生理结构进行分割,其中,对于有左右两部分的生理结构按同一类别进行分割;以及对于所述有左右两部分的生理结构的每一的分割结果,仅保留两个最大的连通域,并基于所述两个最大的连通域的左右相对位置,进一步将其分为所述生理结构的左侧和右侧部分。
Description
本申请总体上涉及一种口腔全景片的分割方法。
如今,牙科诊疗越来越多地借助计算机技术来提高诊疗效率和准确性。
在牙科诊疗中,牙科专业人员常常会用到口腔全景片。口腔全景片又称为曲面断层片,是应用窄缝及圆弧轨道断层摄影原理,通过一次成像,获得摄有全部牙及周围组织总影像的一种简单、快捷的技术。
基于口腔全景片,能够精确测定牙齿倾斜角度,掌握牙周软组织、牙根以及牙槽骨质情况,以及精确测量解剖形态等,为术前诊断分析与治疗方案设计提供全景样图像依据。
一方面,口腔全景片的影像分析要求医生经过长期地系统地实践与学习,因此,经验丰富的医生比较稀缺。另一方面,在临床实践中,医生往往缺乏精力和工具分析和记录全景片提供的信息,如智齿生长发育状况、牙体的破损情况等,这不利于对患者的情况进行进一步评估。
因此,存在对口腔全景片进行自动分析的需求。另外,通过可视化展示的方式有助于医生向患者宣教,促进医生和患者之间的沟通。而要实现口腔全景片的自动分析,首先需要对口腔全景片中的髁突、牙槽骨等生理结构进行分割,因此,有必要提供一种计算机执行的口腔全景片的分割方法。
发明内容
本申请的一方面提供了一种计算机执行的口腔全景片的分割方法,其包括:获取口腔全景片;利用经训练的第一深度神经网络对所述口腔全景片中的生理结构进行分割,其中,对于有左右两部分的生理结构按同一类别进行分割;以及对于所述有左右两部分的生理结构的每一的分割结果,仅保留两个最大的连通域,并基于所述两个最大的连通域的左右相对位置,进一步将其分为所述生理结构的左侧和右侧部分。
在一些实施方式中,所述第一深度神经网络可以是U-Net网络。
在一些实施方式中,所述生理结构包括:上颌窦、髁突、牙槽骨以及下颌神经管,其中,上颌窦、髁突、以及下颌神经管具有左右两部分。
在一些实施方式中,所述的计算机执行的口腔全景片的分割方法还包括:利用经训练的第二深度神经网络自所述口腔全景片中裁剪出目标区域图像,所述利用所述第一深度神经网络的分割是基于所述目标区域图像进行,其中,所述目标区域内包括预先选定的生理结构。
在一些实施方式中,所述第二深度神经网络可以是yolov3。
在一些实施方式中,所述的计算机执行的口腔全景片的分割方法还包括:利用所述经训练的第二深度神经网络检测所述口腔全景片中的第一生理结构区域和第二生理结构区域;以及基于所述检测到的第一生理结构区域和第二生理结构区域的位置关系,以及所述第一生理结构区域或第二生理结构区域的分布,矫正所述目标区域图像的方向,所述利用所述第一深度神经网络的分割是基于所述经矫正方向的目标区域图像进行。
在一些实施方式中,所述第一生理结构区域和第二生理结构区域是以矩形框表达。
在一些实施方式中,所述第一生理结构是牙槽骨,所述第二生理结构是上颌窦,所述方向的矫正是基于所述检测到的牙槽骨区域的长宽方向以及所述牙槽骨区域和上颌窦区域之间的位置关系而进行。
在一些实施方式中,所述的计算机执行的口腔全景片的分割方法还包括:对所述分割结果进行形态学闭运算。
在一些实施方式中,所述的计算机执行的口腔全景片的分割方法还包括:提取所述经形态学闭运算的分割结果的轮廓点;以及基于所述提取得到的轮廓点,以样条曲线拟合得到新的轮廓线,作为最终的分割结果。
以下将结合附图及其详细描述对本申请的上述及其他特征作进一步说明。应当理解的是,这些附图仅示出了根据本申请的若干示例性的实施方式,因此不应被视为是对本申请保护范围的限制。除非特别指出,附图不必是成比例的,并且其中类似的标号表示类似的部件。
图1为本申请一个实施例中的计算机执行的口腔全景片的分割方法的示意性流程图;
图2A为本申请一个实施例中的用于对口腔全景片中进行分割的计算机程序的一个界面所展示的一个例子中的口腔全景片;
图2B为所述计算机程序的一个界面所展示的自图2A所示的口腔全景片裁剪出的目标区域图像;以及
图2C为所述计算机程序的一个界面所展示的图2B所示的目标区域图像的分割结果。
以下的详细描述中引用了构成本说明书一部分的附图。说明书和附图所提及的示意性实施方式仅仅出于是说明性之目的,并非意图限制本申请的保护范围。
在本申请的启示下,本领域技术人员能够理解,可以采用许多其他实施方式,并且可以对所描述实施方式做出各种改变,而不背离本申请的主旨和保护范围。应当理解的是,在此说明并图示的本申请的各个方面可以按照很多不同的配置来布置、替换、组合、分离和设计,这些不同配置都在本申请的保护范围之内。
本申请的一方面提供了一种计算机执行的口腔全景片的分割方法。
请参图1,为本申请一个实施例中的计算机执行的口腔全景片的分割方法100的示意性流程图。
本申请又一方面提供了一种用于对口腔全景片进行分割的计算机系统,其包括存储装置和处理器,所述存储装置存储有一计算机程序,当其被所述处理器运行后,将执行所述口腔全景片的分割方法100。
在101中,获取待处理的口腔全景片。
口腔全景片的拍摄为业界所习知,此处不再赘述。
本申请的口腔全景片的分割方法100不仅能够处理原始拍摄的口腔全景片,还能够处理翻拍的口腔全景片,例如,利用智能移动电话翻拍纸质口腔全景片得到的图片。
请参图2A,为本申请一个实施例中的用于对口腔全景片中进行分割的计算机程序的一个界面所展示的一个例子中的口腔全景片。
在103中,自所述口腔全景片裁剪出目标区域图像。
所述口腔全景片可能是原始拍摄获得,也可能是二次拍摄获得(例如,利用智能移动电话拍摄口腔全景片实物或者计算机屏幕展示的口腔全景片获得)。二次拍摄获得的口腔全景片中可能包括一些不必要的内容,例如,二次拍摄时原始口腔全景片周围的景物,这可能会在一定程度上对后续处理产生影响。
另一方面,对于牙科诊疗而言,只需要口腔全景片中包括预先选定的生理结构的一部分图像,例如,上颌窦、下颌神经管、牙槽骨、髁突以及牙齿。
为了利于后续分割网络的训练,同时提高分割的准确性,可以自所述口腔全景片中裁剪出包含预先选定的生理结构的目标区域的图像,基于裁剪出的目标区域图像进行分割。在一个实施例中,所述目标区域可以是矩形。
在一个实施例中,可以利用经训练的深度神经网络检测口腔全景片中的目标区域,并基于检测的结果裁剪出目标区域图像,以下将该深度神经网络称为目标检测网络。
在一个实施例中,除了目标区域之外,还可以利用所述目标检测网络检测口腔全景片中的上颌窦区域以及牙槽骨区域,并基于检测结果矫正目标区域图像的方向,具体操作如下。
可以理解,由于此处上颌窦区域和牙槽骨区域的检测结果仅用于方向矫正,故对其检测精度要求不高。在一个实施例中,可以用矩形框表达上颌窦区域和牙槽骨区域的检测结果。
由于牙槽骨区域一般为长方形,如果检测出的牙槽骨区域的纵向尺寸大于横向尺寸,则将图片逆时针旋转90度。
上颌窦区域应位于牙槽骨区域上方,如果上颌窦区域位于牙槽骨区域下方,则将图片逆时针旋转180度。
在本申请的启示下,可以理解,除了上颌窦区域和牙槽骨区域之外,也可以采用其他生理结构的组合来矫正目标区域图像的方向,例如,上颌窦区域和下颌神经管区域的组合,或髁突区域和下颌神经管区域的组合等。
在一个实施例中,所述目标检测网络可以采用yolov3网络。可以理解,除了yolov3网络之外,所述目标检测网络也可以采用任何其他适用的深度神经网络,例如,SSD以及Faster RCNN网络等。
在一个实施例中,在训练所述目标检测网络时,可以采用翻转、旋转90度等方法增广训练集,以提高所述目标检测网络的鲁棒性。
请参图2B,为所述计算机程序的一个界面所展示的自图2A所示的口腔全景片裁剪出的目标区域图像。
可以理解,虽然基于目标区域图像进行相关生理结构的分割在网络训练和分割准确性方面具有优势,但直接基于原始的口腔全景片进行分割也是可行的。
在105中,对所述目标区域图像进行生理结构分割。
在一个实施例中,可以利用经训练的深度神经网络对目标区域图像中的生理结构进行分割,以下将该深度神经网络称为分割网络。
在一个实施例中,所述分割网络可以采用U-Net网络。可以理解,除了U-Net网络之外,所述分割网络也可以采用任何其他适用的深度神经网络,例如,DeepLab或FCN等语义分割网络。
在一个实施例中,可以从所述目标区域图像中分割出以下生理结构:牙槽骨、上颌窦、髁突以及下颌神经管。
由于上颌窦、髁突以及下颌神经管具有左右两部分,为了区分左右,对于它们的分割结果,可以采用以下方法进行处理。
对于上颌窦、髁突以及下颌神经管每一的分割结果,计算分割图的连通域,并保留最大的两个连通域。
接着,计算所述两个连通域中心的横坐标,根据该计算的结果,将上颌窦、髁突、下颌神经管进一步分为左侧颌窦、左侧髁突、左侧下颌神经管以及右侧颌窦、右侧髁突、右侧下颌神经管。
在107中,对分割结果进行后处理。
以上获得的分割结果可能存在空洞以及轮廓不够光滑等瑕疵,因此,可以对其进行相应的后处理。
在一个实施例中,可以对以上的分割结果进行形态学闭运算,以消除其中可
能存在的空洞。
接着,提取经形态学闭运算获得的分割结果的轮廓点集,然后,基于提取得到的轮廓点集,以样条曲线拟合得到较为光滑的新的轮廓线,并将此作为最终的分割结果。
请参图2C,为所述计算机程序的一个界面所展示的图2B所示的目标区域图像的分割结果。
可以理解,以上实施例中所列生理结构仅是出于示例性之目的,在实际应用中可以根据具体需求进行增减。
尽管在此公开了本申请的多个方面和实施例,但在本申请的启发下,本申请的其他方面和实施例对于本领域技术人员而言也是显而易见的。在此公开的各个方面和实施例仅用于说明目的,而非限制目的。本申请的保护范围和主旨仅通过后附的权利要求书来确定。
同样,各个图表可以示出所公开的方法和系统的示例性架构或其他配置,其有助于理解可包含在所公开的方法和系统中的特征和功能。要求保护的内容并不限于所示的示例性架构或配置,而所希望的特征可以用各种替代架构和配置来实现。除此之外,对于流程图、功能性描述和方法权利要求,这里所给出的方框顺序不应限于以同样的顺序实施以执行所述功能的各种实施例,除非在上下文中明确指出。
除非另外明确指出,本文中所使用的术语和短语及其变体均应解释为开放式的,而不是限制性的。在一些实例中,诸如“一个或多个”、“至少”、“但不限于”这样的扩展性词汇和短语或者其他类似用语的出现不应理解为在可能没有这种扩展性用语的示例中意图或者需要表示缩窄的情况。
Claims (10)
- 一种计算机执行的口腔全景片的分割方法,其包括:获取口腔全景片;利用经训练的第一深度神经网络对所述口腔全景片中的生理结构进行分割,其中,对于有左右两部分的生理结构按同一类别进行分割;以及对于所述有左右两部分的生理结构的每一的分割结果,仅保留两个最大的连通域,并基于所述两个最大的连通域的左右相对位置,进一步将其分为所述生理结构的左侧和右侧部分。
- 如权利要求1所述的计算机执行的口腔全景片的分割方法,其特征在于,所述第一深度神经网络是U-Net网络。
- 如权利要求1所述的计算机执行的口腔全景片的分割方法,其特征在于,所述生理结构包括:上颌窦、髁突、牙槽骨以及下颌神经管,其中,上颌窦、髁突、以及下颌神经管具有左右两部分。
- 如权利要求1所述的计算机执行的口腔全景片的分割方法,其特征在于,它还包括:利用经训练的第二深度神经网络自所述口腔全景片中裁剪出目标区域图像,所述利用所述第一深度神经网络的分割是基于所述目标区域图像进行,其中,所述目标区域内包括预先选定的生理结构。
- 如权利要求4所述的计算机执行的口腔全景片的分割方法,其特征在于,所述第二深度神经网络是yolov3。
- 如权利要求4所述的计算机执行的口腔全景片的分割方法,其特征在于,它还包括:利用所述经训练的第二深度神经网络检测所述口腔全景片中的第一生理结构区域和第二生理结构区域;以及基于所述检测到的第一生理结构区域和第二生理结构区域的位置关系,以及所述第一生理结构区域或第二生理结构区域的分布,矫正所述目标区域图像的方向,所述利用所述第一深度神经网络的分割是基于所述经矫正方向的目标区域图像进行。
- 如权利要求6所述的计算机执行的口腔全景片的分割方法,其特征在于,所述第一生理结构区域和第二生理结构区域是以矩形框表达。
- 如权利要求6所述的计算机执行的口腔全景片的分割方法,其特征在于,所述第一生理结构是牙槽骨,所述第二生理结构是上颌窦,所述方向的矫正是基于所述检测到的牙槽骨区域的长宽方向以及所述牙槽骨区域和上颌窦区域之间的位置关系而进行。
- 如权利要求1所述的计算机执行的口腔全景片的分割方法,其特征在于,它还包括:对所述分割结果进行形态学闭运算。
- 如权利要求8所述的计算机执行的口腔全景片的分割方法,其特征在于,它还包括:提取所述经形态学闭运算的分割结果的轮廓点;以及基于所述提取得到的轮廓点,以样条曲线拟合得到新的轮廓线,作为最终的分割结果。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210729619.0 | 2022-06-24 | ||
CN202210729619.0A CN117765002A (zh) | 2022-06-24 | 2022-06-24 | 口腔全景片的分割方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2023246463A1 true WO2023246463A1 (zh) | 2023-12-28 |
Family
ID=89379136
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2023/097792 WO2023246463A1 (zh) | 2022-06-24 | 2023-06-01 | 口腔全景片的分割方法 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN117765002A (zh) |
WO (1) | WO2023246463A1 (zh) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109978841A (zh) * | 2019-03-12 | 2019-07-05 | 北京羽医甘蓝信息技术有限公司 | 基于深度学习的全景片阻生牙识别的方法和装置 |
CN113160151A (zh) * | 2021-04-02 | 2021-07-23 | 浙江大学 | 基于深度学习及注意力机制的全景片龋齿深度识别方法 |
WO2021155230A1 (en) * | 2020-01-31 | 2021-08-05 | James R. Glidewell Dental Ceramics, Inc. | Teeth segmentation using neural networks |
CN113223010A (zh) * | 2021-04-22 | 2021-08-06 | 北京大学口腔医学院 | 口腔图像多组织全自动分割的方法和系统 |
CN114332123A (zh) * | 2021-12-30 | 2022-04-12 | 杭州电子科技大学 | 基于全景片的龋病自动分级方法及系统 |
-
2022
- 2022-06-24 CN CN202210729619.0A patent/CN117765002A/zh active Pending
-
2023
- 2023-06-01 WO PCT/CN2023/097792 patent/WO2023246463A1/zh unknown
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109978841A (zh) * | 2019-03-12 | 2019-07-05 | 北京羽医甘蓝信息技术有限公司 | 基于深度学习的全景片阻生牙识别的方法和装置 |
WO2021155230A1 (en) * | 2020-01-31 | 2021-08-05 | James R. Glidewell Dental Ceramics, Inc. | Teeth segmentation using neural networks |
CN113160151A (zh) * | 2021-04-02 | 2021-07-23 | 浙江大学 | 基于深度学习及注意力机制的全景片龋齿深度识别方法 |
CN113223010A (zh) * | 2021-04-22 | 2021-08-06 | 北京大学口腔医学院 | 口腔图像多组织全自动分割的方法和系统 |
CN114332123A (zh) * | 2021-12-30 | 2022-04-12 | 杭州电子科技大学 | 基于全景片的龋病自动分级方法及系统 |
Also Published As
Publication number | Publication date |
---|---|
CN117765002A (zh) | 2024-03-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Jader et al. | Deep instance segmentation of teeth in panoramic X-ray images | |
Leite et al. | Artificial intelligence-driven novel tool for tooth detection and segmentation on panoramic radiographs | |
Wirtz et al. | Automatic teeth segmentation in panoramic X-ray images using a coupled shape model in combination with a neural network | |
JP6483273B2 (ja) | 口腔内画像の自動選択及びロック | |
Yüksel et al. | Dental enumeration and multiple treatment detection on panoramic X-rays using deep learning | |
US8711178B2 (en) | System and method for generating profile morphing using cephalometric tracing data | |
JP6830082B2 (ja) | 歯科分析システムおよび歯科分析x線システム | |
Leonardi et al. | Fully automatic segmentation of sinonasal cavity and pharyngeal airway based on convolutional neural networks | |
Oliveira et al. | Caries detection in panoramic dental X-ray images | |
US11864936B2 (en) | Systems and methods for determining orthodontic treatment | |
WO2020215810A1 (zh) | 一种基于图像识别的结肠镜手术中窄带成像检测方法 | |
US12062170B2 (en) | System and method for classifying a tooth condition based on landmarked anthropomorphic measurements | |
US20220358740A1 (en) | System and Method for Alignment of Volumetric and Surface Scan Images | |
Chen et al. | Automated Detection System Based on Convolution Neural Networks for Retained Root, Endodontic Treated Teeth, and Implant Recognition on Dental Panoramic Images | |
US20220222814A1 (en) | System and method for facial and dental photography, landmark detection and mouth design generation | |
Li et al. | Automated integration of facial and intra-oral images of anterior teeth | |
WO2023246462A1 (zh) | 口腔全景片中恒牙和乳牙的分割和牙号确定方法 | |
Chen et al. | Detection of various dental conditions on dental panoramic radiography using Faster R-CNN | |
CN107802276B (zh) | 头颅影像的描迹图绘制装置和方法 | |
Jiao et al. | Deep learning for automatic detection of cephalometric landmarks on lateral cephalometric radiographs using the Mask Region-based Convolutional Neural Network: a pilot study | |
WO2023246463A1 (zh) | 口腔全景片的分割方法 | |
US20230013902A1 (en) | System and Method for Correcting for Distortions of a Diagnostic Image | |
Cunha et al. | A method for segmentation of dental implants and crestal bone | |
CN115908361A (zh) | 一种口腔全景片的龋齿识别方法 | |
İçöz et al. | Evaluation of an Artificial Intelligence System for the Diagnosis of Apical Periodontitis on Digital Panoramic Images |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23826112 Country of ref document: EP Kind code of ref document: A1 |