CN111354004B - A segmentation method of left and right ear regions based on temporal bone CT plain scan images - Google Patents

A segmentation method of left and right ear regions based on temporal bone CT plain scan images Download PDF

Info

Publication number
CN111354004B
CN111354004B CN202010124604.2A CN202010124604A CN111354004B CN 111354004 B CN111354004 B CN 111354004B CN 202010124604 A CN202010124604 A CN 202010124604A CN 111354004 B CN111354004 B CN 111354004B
Authority
CN
China
Prior art keywords
area
right ear
value
image
temporal bone
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010124604.2A
Other languages
Chinese (zh)
Other versions
CN111354004A (en
Inventor
王云峰
颜波
李健
李吉春
谭伟敏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Fudan University
Original Assignee
Fudan University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Fudan University filed Critical Fudan University
Priority to CN202010124604.2A priority Critical patent/CN111354004B/en
Publication of CN111354004A publication Critical patent/CN111354004A/en
Application granted granted Critical
Publication of CN111354004B publication Critical patent/CN111354004B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10081—Computed x-ray tomography [CT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20112—Image segmentation details
    • G06T2207/20132—Image cropping
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30008—Bone

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Apparatus For Radiation Diagnosis (AREA)
  • Image Analysis (AREA)

Abstract

本发明属于医疗图像处理技术领域，具体为一种基于颞骨CT平扫影像的左右耳区域分割方法。该算法用于对头部颞骨CT影像中的左、右耳部区域进行分割，使得计算机能够集中地对耳部区域进行后续的处理。本发明依次包括以下步骤：对头部CT图像的无效区域进行剔除；按照比例将图像中的左、右耳区域分别裁出；将左、右耳区图像分别整理保存。实验结果表明，对于平扫的颞骨CT图像，本方法可以剔除无效的黑色区域，抛弃其他无关部位的信息，准确地提取左右耳区域。

Figure 202010124604

The invention belongs to the technical field of medical image processing, in particular to a left and right ear region segmentation method based on a temporal bone CT plain scan image. The algorithm is used to segment the left and right ear regions in the CT image of the head temporal bone, so that the computer can focus on the subsequent processing of the ear region. The invention includes the following steps in sequence: removing the invalid area of the head CT image; cutting out the left and right ear areas in the image according to the proportion; sorting and saving the left and right ear area images respectively. The experimental results show that for the plain-scanned temporal bone CT images, this method can eliminate the invalid black areas, discard the information of other irrelevant parts, and accurately extract the left and right ear regions.

Figure 202010124604

Description

Left and right ear region segmentation method based on temporal bone CT flat scan image
Technical Field
The invention belongs to the technical field of medical image processing, and particularly relates to a left and right ear region segmentation method based on a temporal bone CT flat scan image.
Background
With the development of computer science, intelligent medical treatment becomes a great technological innovation for improving the modern medical treatment level. The advantages it has in various aspects are gaining increasing acceptance and attention. Among them, the medical image processing technology is one of the key parts for promoting the development of intelligent medical treatment, and is also a very important application link of intelligent medical treatment.
In the application scene of otology CT image diagnosis, doctors can only obtain a whole temporal bone CT image at the present stage, so as to diagnose the ear pathological changes of patients. However, when the doctor actually makes a diagnosis, only the ear region in the temporal bone CT image is observed and relied on. The area is very small, and the length and width of a single area are less than 1/5 of the whole picture. Other redundant CT information is redundant without assistance for otology diagnosis. Not only doctors, but also otology image processing algorithms, redundant non-ear information is useless for the operation and judgment of ear information, and sometimes even serious interference is generated. Therefore, a program or an algorithm is needed to process the temporal bone CT image and extract the left and right ear regions.
So far, no special algorithm is used for segmenting and extracting left and right ear regions of a temporal bone CT image.
Disclosure of Invention
In order to fill up the blank of related algorithms, the invention provides a left and right ear region segmentation method based on a temporal bone CT flat scan image.
The invention provides a left and right ear region segmentation method based on a temporal bone CT flat scan image.
(1) Cropping and size normalization of active areas
Suppose an input temporal bone CT medical image I of size H0×W0(ii) a Wherein the gray value of each pixel has a value range of [0,65535](ii) a The method adopts T as a threshold value binarization input image to obtain a mapping Map, so that an effective area is 1, an invalid black area is 0, and a formula is recorded as follows:
Figure BDA0002394043780000011
wherein, the range of the T can be taken as [400,550], and the value is suggested to be 500 according to the value range of the pixel gray value and the gray value distribution of the effective area; and then, carrying out transverse and longitudinal summation on the Map to obtain MapX and MapY, wherein the formula is recorded as:
Figure BDA0002394043780000012
Figure BDA0002394043780000013
using MapY meterCalculating an upper boundary UP and a lower boundary DOWN of the effective region; UP being greater than threshold T in vector MapY1And DOWN is the index of the first element of the vector MapY above the threshold T1Subscript of the last element of (a); wherein, T1The value is 10, namely, when the number of the effective pixels in the horizontal direction reaches more than 10, the effective pixels are regarded as the boundary; in the same way, MapX is used to find the LEFT boundary LEFT and the RIGHT boundary RIGHT of the effective area, where the threshold is also T1；
Cutting I into I', and expressing the formula as follows:
I′＝{I(x,y)|LEFT≤x≤RIGHT,UP≤y≤DOWN} (4)
finally, unifying the size of I' to H by using a Bicubic interpolation method1×W1(ii) a Here H1、W1May be different from the original size H0、W0The same may be true; for simple operation, uniform and moderate input size, H1、W1A suggested value is 512.
(2) Cropping of left and right ear regions and discarding of extraneous regions
In the normalized effective region CT image I', the Area of the left ear is croppedlArea of right earrThe method comprises the following steps:
Areal＝{I′(x,y)|leftl≤x≤rightl,up≤y≤down} (5)
Arear＝{I′(x,y)|leftr≤x≤rightr,up≤y≤down} (6)
left of the above typel,rightlUp, down are boundary variables of the left and right ear regions; wherein, according to the characteristics of the CT flat scanning image of the temporal bone, leftlThe value range is [260,265 ]]，rightlThe value range is [510,512 ]]，leftrThe value range is [1,5 ]]，rightrThe value range is [245,255 ]]And up has a value range of [195,205 ]]The down value range is [325,335 ]]。
(3) Classification and arrangement of left and right ear regions
For temporal bone CT medical image set to be processed
Figure BDA0002394043780000021
Take picture I thereinjRepeating the steps (2) and (3), and combining the final results to form a final left ear CT area set
Figure BDA0002394043780000022
And right ear CT region set
Figure BDA0002394043780000023
The formula is expressed as:
Figure BDA0002394043780000024
Figure BDA0002394043780000025
the invention has the beneficial effects that: the left and right ear images are clearly and completely extracted from the whole temporal bone CT image, the operation speed is high, and the method is simple and effective. Under the application scene of otology disease diagnosis, the independent left and right ear CT images can better serve doctors or intelligent medical systems, effective information is provided, and the interference of redundant information is avoided.
Drawings
FIG. 1 is a flow chart of the present invention.
Figure 2 is a demonstration of the results using the present invention.
Detailed Description
For an input temporal bone CT picture I, the left and right ear areas are segmented and extracted. The method comprises the following specific steps:
(1) and (5) carrying out binarization on the effective region of the I by using a threshold value T of 500 to obtain a binarized feature map. And calculating an upper boundary UP, a lower boundary DOWN, a LEFT boundary LEFT and a RIGHT boundary RIGHT of the effective region according to the feature map. Cutting the I according to the boundary, and utilizing Bicubic algorithm interpolation to enlarge the cut effective area CT image to 512 multiplied by 512;
(2) cropping of left and right ear regions is associated with discarding of extraneous regions.The range of the right ear region is: upper boundary up of 200, lower boundary down of 300, left boundary leftrGet 1, right margin rightrTaking 250; the range of the left ear region is: upper boundary up of 200, lower boundary down of 300, left boundary leftlGet 263, right border rightlAnd taking 512. According to the range, cutting the two regions in the CT image of the effective region to obtain the left and right ear regions of a single picture;
(3) and (5) classifying and sorting the left ear region and the right ear region. The left ear and right ear regions of all pictures are repeatedly calculated and divided, and the results are merged to finally form a left ear CT image set and a right ear CT image set.
FIG. 2 is a graph showing the results of the present invention. It can be seen that the invention can correctly segment the left and right ear regions from the entire temporal bone CT.

Claims (7)

1.一种基于颞骨CT平扫影像的左右耳区域分割方法，其特征在于，具体步骤如下：1. a left and right ear region segmentation method based on temporal bone CT plain scan image, is characterized in that, concrete steps are as follows: (1)有效区域的裁剪与尺寸标准化(1) Cutting and size standardization of the effective area 假定输入的颞骨CT医疗图像I，其尺寸为H0×W0，其像素灰度值取值范围为[0，65535]，采用T作为阈值二值化输入图，得到映射图Map，使得有效区域为1，无效黑色区域为0，公式记作：Assume that the input temporal bone CT medical image I, its size is H 0 ×W 0 , its pixel gray value range is [0, 65535], and T is used as the threshold to binarize the input image to obtain the map Map, which makes the effective The area is 1, the invalid black area is 0, and the formula is written as:
Figure FDA0003277544050000011
Figure FDA0003277544050000011
 对Map进行纵向与横向求和，得到MapX与MapY，公式记作：Sum the Map vertically and horizontally to get MapX and MapY. The formula is written as:
Figure FDA0003277544050000012
Figure FDA0003277544050000012
Figure FDA0003277544050000013
Figure FDA0003277544050000013
 利用MapY计算有效区域的上边界UP，下边界DOWN，UP为向量MapY中大于阈值T1的第一个元素的下标，而DOWN为向量MapY中阈值大于T1的最后一个元素的下标；同理利用MapX分别求得有效区域的左边界LEFT，右边界RIGHT；Use MapY to calculate the upper boundary UP and lower boundary DOWN of the effective area, UP is the subscript of the first element in the vector MapY that is greater than the threshold T1, and DOWN is the subscript of the last element in the vector MapY whose threshold is greater than T1; Similarly, use MapX to obtain the left boundary LEFT and the right boundary RIGHT of the effective area respectively; 将I裁剪为I’，公式表达如下：Cut I to I', the formula is as follows: I′＝{I(x，y)|LEFT≤x≤RIGHT，UP≤y≤DOWN} (4)I′={I(x, y)|LEFT≤x≤RIGHT, UP≤y≤DOWN} (4) 最后，再使用Bicubic插值方法，将I’的尺寸统一到H1×W1；Finally, use the Bicubic interpolation method to unify the size of I' to H 1 ×W 1 ; (2)左右耳区域的裁剪与无关区域的丢弃(2) Cropping of left and right ear regions and discarding of irrelevant regions 在标准化有效区域CT图像I’中，裁切左耳区域Areal，右耳区域Arear如下：In the normalized effective area CT image I', the left ear area Area l and the right ear area Area r are cropped as follows: Areal＝{I′(x，y)|leftl≤x≤rightl，up≤y≤down} (5)Area l ={I′(x, y)|left l ≤x≤right l , up≤y≤down} (5) Arear＝{I′(x，y)|leftr≤x≤rightr，up≤y≤down} (6)Area r ={I′(x, y)|left r ≤x≤right r , up≤y≤down} (6) 其中leftl，rightl，up，down为左右耳区域的边界变量；where left l , right l , up, down are the boundary variables of the left and right ear regions; (3)左右耳区域的归类整理(3) Classification of left and right ear regions 对于待处理的颞骨CT医疗图像集合
Figure FDA0003277544050000014
其中N是CT图像集合中元素的个数；取遍其中图片Ij并重复步骤(1)、(2)、(3)，将其最终结果进行合并，形成最终的左耳CT区域集合
Figure FDA0003277544050000015
与右耳CT区域集合
Figure FDA0003277544050000016
公式表达为：For the collection of temporal bone CT medical images to be processed
Figure FDA0003277544050000014
Wherein N is the number of elements in the CT image set; take all the pictures I j and repeat steps (1), (2), (3), and combine the final results to form the final left ear CT area set
Figure FDA0003277544050000015
Collection with right ear CT area
Figure FDA0003277544050000016
The formula is expressed as:
Figure FDA0003277544050000017
Figure FDA0003277544050000017
Figure FDA0003277544050000018
Figure FDA0003277544050000018
 2.根据权利要求1所述的方法，其特征在于，步骤(1)中，阈值T的取值为400-550。2 . The method according to claim 1 , wherein, in step (1), the value of the threshold T is 400-550. 3 . 3.根据权利要求1所述的方法，其特征在于，步骤(1)中，阈值T1的取值为10。3 . The method according to claim 1 , wherein, in step (1), the value of the threshold T 1 is 10. 4 . 4.根据权利要求1所述的方法，其特征在于，步骤(1)中，H1、W1的取值与H0、W0相同。4 . The method according to claim 1 , wherein, in step (1), the values of H 1 and W 1 are the same as H 0 and W 0 . 5.根据权利要求1所述的方法，其特征在于，步骤(2)中，leftl取值为260-265，rightl取值为510-512。5. method according to claim 1 is characterized in that, in step (2), the value of left l is 260-265, and the value of right l is 510-512. 6.根据权利要求1所述的方法，其特征在于，步骤(2)中，leftr取值为1-5，rightr取值为245-255。6 . The method according to claim 1 , wherein in step (2), the value of left r is 1-5, and the value of right r is 245-255. 7 . 7.根据权利要求1所述的方法，其特征在于，步骤(2)中，up取值为195-205，down取值为325-335。7 . The method according to claim 1 , wherein, in step (2), the value of up is 195-205, and the value of down is 325-335. 8 .
CN202010124604.2A 2020-02-27 2020-02-27 A segmentation method of left and right ear regions based on temporal bone CT plain scan images Active CN111354004B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010124604.2A CN111354004B (en) 2020-02-27 2020-02-27 A segmentation method of left and right ear regions based on temporal bone CT plain scan images

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010124604.2A CN111354004B (en) 2020-02-27 2020-02-27 A segmentation method of left and right ear regions based on temporal bone CT plain scan images

Publications (2)

Publication Number Publication Date
CN111354004A CN111354004A (en) 2020-06-30
CN111354004B true CN111354004B (en) 2022-03-18

Family

ID=71192473

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010124604.2A Active CN111354004B (en) 2020-02-27 2020-02-27 A segmentation method of left and right ear regions based on temporal bone CT plain scan images

Country Status (1)

Country Link
CN (1) CN111354004B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1658225A (en) * 2005-03-16 2005-08-24 沈阳工业大学 A Personal Identification Method Based on Geometric Parameters of Auricle
CN102096900A (en) * 2007-08-30 2011-06-15 精工爱普生株式会社 Image processing device, image processing method, and image processing program
EP2840551A1 (en) * 2013-08-23 2015-02-25 Vistaprint Schweiz GmbH Methods and systems for automated selection of regions of an image for secondary finishing and generation of mask image of same
CN104637056A (en) * 2015-02-02 2015-05-20 复旦大学 Method for segmenting adrenal tumor of medical CT (computed tomography) image based on sparse representation

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1658225A (en) * 2005-03-16 2005-08-24 沈阳工业大学 A Personal Identification Method Based on Geometric Parameters of Auricle
CN102096900A (en) * 2007-08-30 2011-06-15 精工爱普生株式会社 Image processing device, image processing method, and image processing program
EP2840551A1 (en) * 2013-08-23 2015-02-25 Vistaprint Schweiz GmbH Methods and systems for automated selection of regions of an image for secondary finishing and generation of mask image of same
CN104637056A (en) * 2015-02-02 2015-05-20 复旦大学 Method for segmenting adrenal tumor of medical CT (computed tomography) image based on sparse representation

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
A Survey on Ear Biometrics;AYMAN ABAZA et al.;《ACM Computing Surveys》;20130228;第1-35页 *
一种基于改进GVF Snake的自动人耳检测方法;李一波等;《模式识别与人工智能》;20100831;第557-559页 *

Also Published As

Publication number Publication date
CN111354004A (en) 2020-06-30

Similar Documents

Publication Publication Date Title
WO2022063199A1 (en) Pulmonary nodule automatic detection method, apparatus and computer system
CN103455991B (en) A kind of multi-focus image fusing method
CN104036479B (en) Multi-focus image fusion method based on non-negative matrix factorization
CN108510489B (en) Pneumoconiosis detection method and system based on deep learning
WO1995014966A1 (en) Automated method and system for the segmentation of medical images
CN110838100A (en) A system for screening and segmentation of colonoscopy pathological sections based on sliding window
CN108830149A (en) A kind of detection method and terminal device of target bacteria
CN110766670A (en) Mammary gland molybdenum target image tumor localization algorithm based on deep convolutional neural network
CN110136161A (en) Image feature extraction and analysis method, system and device
CN107174232B (en) Electrocardiogram waveform extraction method
CN114049339B (en) Fetal cerebellum ultrasonic image segmentation method based on convolutional neural network
CN112053325A (en) Breast mass image processing and classifying system
WO2021139447A1 (en) Abnormal cervical cell detection apparatus and method
CN112052854B (en) A reversible information hiding method for medical images that achieves adaptive contrast enhancement
CN115423806B (en) Breast mass detection method based on multi-scale cross-path feature fusion
CN111340773B (en) A Retinal Image Vessel Segmentation Method
CN114693672B (en) A method for removing skin glands and nipples from mammary gland mammography images based on image processing
CN115272647A (en) Lung image recognition processing method and system
CN111354004B (en) A segmentation method of left and right ear regions based on temporal bone CT plain scan images
CN109816665B (en) A method and device for fast segmentation of optical coherence tomography images
CN117911716B (en) Arthritis CT image feature extraction method based on machine vision
CN112381839B (en) A deep learning-based method for segmentation of HE cancer nests in breast cancer pathological images
CN106446923A (en) Medical image classification method based on corner matching
CN117576119B (en) A semi-supervised left atrium segmentation method based on compression-excitation network
Cyriac et al. Renal calculi detection using image processing

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant