CN110991339B - Three-dimensional palate wrinkle identification method adopting cyclic frequency spectrum - Google Patents
Three-dimensional palate wrinkle identification method adopting cyclic frequency spectrum Download PDFInfo
- Publication number
- CN110991339B CN110991339B CN201911216905.1A CN201911216905A CN110991339B CN 110991339 B CN110991339 B CN 110991339B CN 201911216905 A CN201911216905 A CN 201911216905A CN 110991339 B CN110991339 B CN 110991339B
- Authority
- CN
- China
- Prior art keywords
- palate
- dimensional
- crease
- wrinkle
- spectrum
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the technical field of identity identification of human body biological characteristics legal doctors, and particularly relates to the following technical scheme: a three-dimensional palate wrinkle identification method adopting a cyclic frequency spectrum comprises the following specific steps: acquiring three-dimensional palate wrinkling information by adopting a portable color intraoral scanner; equidistant slicing is carried out on the acquired three-dimensional palate crease image; corroding and binarizing the palate crease slice image; constructing a palate crease slice database; extracting cyclic spectrum characteristics of palate crease lines in slice images; dividing the extracted palate crease line circulation spectrum characteristic diagram by adopting a 4 multiplied by 4 uniform dividing scheme, and further constructing a palate crease line circulation spectrum characteristic vector; finally, three-dimensional palate crease identification is carried out; the equidistant slicing concept provided by the method simplifies the complexity of the three-dimensional palate wrinkle image recognition problem, can reflect the essential characteristics of palate wrinkle tangent lines, avoids the occurrence of data disaster problem by blocking the cyclic spectrum characteristics, can reduce the complexity of calculating the three-dimensional palate wrinkle recognition problem, and accelerates the operation speed.
Description
Technical Field
The invention belongs to the technical field of legal medical expert identification of human body biological characteristics, and particularly relates to a three-dimensional palate crease identification method which is used for legal medical expert identification and adopts equidistant slicing and cyclic spectrum characteristic extraction.
Background
The palate crease is an irregular, asymmetric ridge line located in the anterior third of the hard palate at the top of the human mouth extending outwardly from the anterior incisor mastoid and the palate midseam. Biometric identification is a technique that uses physiological or behavioral characteristics of a person to identify the person. As a new biometric feature, palate wrinkles are equivalent to the role of fingerprints, faces, irises, etc. in biometric feature recognition. The palate wrinkles of each person are unique, and the shape of the palate wrinkles is kept unchanged during the life of the person, and only the length and the width of the palate wrinkles are changed due to normal growth and development. Palate wrinkles meet the requirements of biological characteristics on genetics, stability, acquirability and uniqueness.
Biometric technology plays a very important role in a number of major disasters and criminal cases (e.g., major aviation accidents, natural disasters, industrial explosions, etc.). Because fingerprints, faces, irises and the like are easy to be influenced by factors such as temperature, humidity, huge destructive power and the like, the fingerprints, the faces, the irises and the like are difficult to be completely stored, and certain difficulty is brought to personal identification. The palate is protected from damage in the event of trauma and high temperatures by the presence of soft and hard tissues within the mouth (e.g., lips, cheeks, teeth, bones, etc.). Studies show that palate wrinkles can resist damage such as third degree burn, and the human body has strong anti-putrefaction capability within 7 days of death, and can still keep a more complete form. Thus, the palate crease can be used as a biometric feature for personal identification.
At present, researches on palate wrinkles are mainly focused on fields of mankind, genetics, forensic stomatology, forensic dentistry, anatomies and the like, and mainly classified researches on the forms, the number, the length, the sex, the race and the like of the palate wrinkles in statistical significance are carried out, but researches on three-dimensional digital image recognition are less, and a complete theoretical system is not formed. In 1732, winslow first proposed the anatomical concept of "palate wrinkling". In 1889, allen et al first used palate wrinkles as a tool for identification during the study of palate wrinkles. In 1955, lysell developed the first palate wrinkle classification system. Subsequently, a number of researchers classified the shape, number, length, orientation, etc. of the palate wrinkles, inventing different types of palate wrinkle classification systems and using them for personal identification, with the classification methods of Kapali and Thomas et al being most commonly used. In 2007, several researchers have systematically summarized and arranged the most commonly used classification methods, such as the Gamea classification method and the Martins dos Santos classification method, based on the past palate crease classification. In 2010, hamanth M used software MS paint version5.1 to extract and match information from the palate wrinkling image. In 2018 Bernitz et al conducted the same-identification study on 1 forensic sample using the radiation transformation principle, again underscores the advantage of palate wrinkling as a discrimination tool when DNA or the like is not available for the same-identification study. Compared with foreign research work on palate wrinkle identification, domestic research on palate wrinkle identification starts later. In 2015, pan Fei et al discussed that palate wrinkles have three characteristics of stability, variability and versatility required by forensic identity. In 2016, wu Xiuping et al established and evaluated the efficacy of the same digital system as in the palate-wrinkled forensic science. Gu Yu discusses the importance of palate wrinkling in open-mouth forensics 2017. Li Bing demonstrates the feasibility of forensic stomatology identity recognition using digital images of the palate wrinkles. In general, the study of students at home and abroad on two-dimensional palate wrinkling images is mainly focused on extracting or counting the shape, the number and the length of palate wrinkling stripes and the edge characteristics of the stripes.
So far, little research has been done on three-dimensional palate crease digital image recognition. The traditional method collects three-dimensional palate wrinkling data by making a gypsum model, and has the problems of easy damage, time-consuming collection, difficult storage and the like. In 2015, taniva E D made the palate crease into a plaster model, then scanned the model into a three-dimensional palate crease digital image by a scanning instrument, and finally used for characteristic matching of the image. In 2016, batool Ali et al used Thamas and Kotze classification to evaluate the length and shape of the palate wrinkles in a three-dimensional palate wrinkles plaster model after orthodontic treatment of a patient. In 2017 Daniele Gibeli et al verified the uniqueness of the three-dimensional model of the palate crease and applied it for authentication. In the same year, wu Xiao snow is used for photographing the palate wrinkling plaster model by using a digital camera, coding the number of palate wrinkling strips, the duty ratio of the palate wrinkling strips and the like, and finally researching feature matching. In 2019, gu Yu developed a study on the correlation of the palate wrinkles, the gender and the blood type by using a three-dimensional digital palate wrinkles model, and explored the possibility of identifying the gender by using the length and the shape of the palate wrinkles.
Disclosure of Invention
The invention provides a method for identifying human biological characteristics by legal medical expert identification, which can be used for completing three-dimensional palate wrinkle identification and has the advantages of accuracy, rapidness and good identification effect.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: a three-dimensional palate wrinkle identification method adopting a cyclic frequency spectrum specifically comprises the following steps:
step one, collecting three-dimensional palate wrinkling information. Specifically, the collecting device is a portable color intraoral scanner (comprising a scanning gun, a USB secret key, a software USB flash disk, a POD base, a color calibration head and a notebook computer), the position of a dental chair is adjusted to be in a position where the lower jaw of a volunteer is parallel to the ground and the upper jaw of the volunteer is perpendicular to the ground, the volunteer and the collecting device are respectively sterilized, then the gun head of the scanning gun is stretched into the oral cavity of the volunteer, the scanning is carried out along teeth in a moving way, whether all surfaces of the palate wrinkles are complete or not is checked while the scanning is carried out, if the loophole exists or the scanning is incomplete, the supplementing scanning is carried out, and after the scanning is finished, data are imported into the computer to obtain three-dimensional palate wrinkle data.
And step two, equidistant slicing is carried out on the palate crease image. The slicing principle is as follows: the three-dimensional palate crease model is intersected with a horizontal slice plane to carry out slicing, and equidistant slicing refers to slicing which generates constant spacing and constant thickness. The specific operation of slicing is as follows: (1) Firstly, adopting 3Shape software to open a three-dimensional palate wrinkling image, measuring the length, width and height of the palate wrinkling image by using a rectangular frame of the software, and preparing uniform slices of the three-dimensional palate wrinkling image by using a circular cutting line; (2) Firstly, adjusting the position of a rectangular frame to enable the upper surface and the lower surface of a palate crease image to be overlapped (namely, the upper surface and the lower surface of the rectangular frame are overlapped), connecting the protruding points of the three teeth from the left tooth and the right tooth of the incisor mastoid in a top view to form a cutting straight line, then adopting a circular cutting line to start cutting from the left tooth and the right tooth of the incisor mastoid, continuously adjusting the positions of the circular cutting line and the rectangular frame before starting equidistant slicing to enable the circular cutting line to be parallel to the cutting straight line, and finally, equidistant slicing along the midpoint of the circular cutting line and the incisor mastoid direction (the vertical line is the slicing direction, the vertical line grows between 39cm and 40cm from bottom to top, and the vertical line is perpendicular to the circular cutting line); (3) The constant spacing adopted by the invention is 1cm, equidistant slicing operation is carried out on three-dimensional palate wrinkles once every 1cm, the three-dimensional palate wrinkles are sequentially operated for 40 times, 40 two-dimensional palate wrinkles slice images are sequentially obtained after equidistant slicing, and are stored in jpg format, and the 40 slice images together form a palate wrinkles slice sample set.
And thirdly, constructing a jaw wrinkle section database, wherein the jaw wrinkle section image database is built according to the unique characteristics of the jaw wrinkle images, 5 jaw wrinkle samples are acquired for 91 persons, the equidistant section operation of the second step is carried out on each jaw wrinkle sample, and then 40 jaw wrinkle equidistant section curve images of each jaw wrinkle sample can be acquired, so that 455 jaw wrinkle samples and 18200 Zhang Ezhou section images of 91 persons can be obtained. Wherein each person's 5 palate-wrinkled samples are three-dimensional palate-wrinkled images taken at different times, with the same light dark background and with a slight rotation.
And fourthly, corroding and binarizing the palate crease slice image, in order to eliminate irrelevant information in the image, recovering useful real information, enhancing the detectability of relevant information and simplifying data to the greatest extent, thereby enhancing the reliability of palate crease feature extraction, palate crease image segmentation, palate crease matching and identification.
Step five, extracting the cyclic spectrum characteristics of the palate crease line in the slice image, constructing a palate crease characteristic vector by extracting the cyclic spectrum characteristics of the palate crease line, specifically, firstly scanning the palate crease line image to obtain one-dimensional components in the horizontal direction and the vertical direction, and further adopting cumulative Fourier transformation to respectively calculate spectrum correlation functions of the two one-dimensional components.
And step six, constructing a jaw wrinkle tangent line circulation spectrum characteristic vector, wherein in order to avoid dimension disasters and facilitate subsequent identification processing, the method adopts a 4 multiplied by 4 uniform block scheme to block the jaw wrinkle tangent line circulation spectrum characteristic diagram extracted in the step five. The blocking principle is as follows: and uniformly partitioning based on pixel positions, solving the mean value and the variance of each small block, and finally converting the mean value and the variance into one-dimensional row vectors in sequence.
And seventhly, three-dimensional palate wrinkles are identified, wherein the three-dimensional palate wrinkles identification system adopted by the method comprises two components of a three-dimensional palate wrinkles image training process and a three-dimensional palate wrinkles image testing process, and each component comprises a three-dimensional palate wrinkles preprocessing module, a palate wrinkles tangent line circulation frequency spectrum characteristic extraction module, a palate wrinkles tangent line circulation frequency spectrum characteristic segmentation module and a palate wrinkles tangent line characteristic vector classification module.
The equidistant slicing concept provided by the method simplifies the complexity of the three-dimensional palate wrinkle image recognition problem, adopts the cyclic spectrum characteristics to construct the characteristic vector, can reflect the essential characteristics of palate wrinkle tangent lines, avoids the occurrence of data disaster problem by the blocking processing of the cyclic spectrum characteristics, and can reduce the complexity of calculating the three-dimensional palate wrinkle recognition problem and accelerate the operation in the actual operation process.
Drawings
Fig. 1 is a schematic representation of a three-dimensional digital image of palate crease.
Fig. 2 is a schematic view of a circular cut line of the palate crease in the 3Shape software.
Fig. 3 is a schematic diagram of a standard top view of a three-dimensional palate crease slice.
Fig. 4 is a schematic diagram of a three-dimensional palate crease equidistant slicing process.
Fig. 5 is a schematic representation of a slice sample set of palate wrinkles.
FIG. 6 is a schematic diagram of the erosion and binarization process of the palate crease line image.
Fig. 7 is a schematic diagram of a palate crease tangent cycle spectrum feature extraction process.
Fig. 8 is a schematic diagram of a three-dimensional palate crease identification system.
Fig. 9 is a block diagram of a palate crease tangent cycle spectrum feature.
Fig. 10 is a schematic diagram of the construction process of the palate crease tangent cycle spectrum characteristic vector.
Fig. 11 is a schematic diagram of a three-dimensional palate crease identification system.
Detailed Description
In order to make the technical problems, technical schemes and beneficial effects to be solved more clear, the invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
A three-dimensional palate wrinkle identification method adopting a cyclic frequency spectrum comprises the following specific steps:
step one, collecting three-dimensional palate wrinkling information. Specifically, the collecting device is a portable color intraoral scanner (comprising a scanning gun, a USB secret key, a software U disk, a POD base, a color calibration head and a notebook computer), the position of a dental chair is adjusted to be in a position where the lower jaw of a volunteer is parallel to the ground and the upper jaw of the volunteer is perpendicular to the ground, the volunteer and the collecting device are respectively disinfected, then the gun head of the scanning gun is stretched into the oral cavity of the volunteer, the scanning is carried out along teeth in a moving way, whether all surfaces of the palate wrinkles are complete or not is checked while the scanning is carried out, if the loophole is present or the scanning is incomplete, the supplementing scanning is carried out, after the scanning is completed, data is imported into the computer to obtain three-dimensional palate wrinkle data, as shown in fig. 1, compared with the traditional two-dimensional palate wrinkle image, the three-dimensional palate wrinkle image can reflect the distribution of the palate wrinkles more three-dimensionally, intuitively and in detail, the interference caused by lips, saliva, light and the like can be avoided, the palate wrinkle image obtained after the oral wrinkle image scanned by the oral cavity scanner is more in accordance with the real distribution condition of the oral cavity of the palate wrinkles of a human body, the palate wrinkles is arranged in order, and detail information can be kept more clearly.
And step two, equidistant slicing is carried out on the palate crease image. Firstly, the complexity of the problem is reduced, the three-dimensional palate crease image is subjected to equidistant slicing treatment, and the slicing principle is as follows: the three-dimensional palate model is sliced intersecting horizontal slice planes, each giving a piecewise linear contour of the slice, equidistant slices referring to the resulting slices that produce constant spacing and a fixed constant thickness. The specific operation of slicing is as follows: (1) Firstly, a three-dimensional palate wrinkling image is opened by adopting 3Shape software, the length, width and height of the palate wrinkling image are measured by using a rectangular frame of the software, and a circular cutting line (shown in figure 2) is used for preparing uniform slices of the three-dimensional palate wrinkling image; (2) As shown in figure 3, the positions of the rectangular frame are firstly adjusted to enable the upper surface and the lower surface of the palate crease image to be overlapped (namely, the upper surface and the lower surface of the rectangular frame are overlapped), in a top view, three teeth are respectively arranged from the left side and the right side of the incisor mastoid, the protruding points of the three teeth are connected into a cutting straight line, then a circular cutting line is adopted to cut from the five teeth on the left side and the right side of the incisor mastoid, before equidistant slicing is started, the positions of the circular cutting line and the rectangular frame are continuously adjusted to enable the circular cutting line to be parallel to the cutting straight line, and finally equidistant slicing is carried out along the midpoint of the circular cutting line and the incisor mastoid direction (note: the vertical line is the slicing direction, the vertical line is approximately 39cm to 40cm from bottom to top, and the vertical line is perpendicular to the circular cutting line); (3) As shown in fig. 4, the constant spacing adopted by the invention is 1cm, equidistant slicing operation is performed on three-dimensional palate wrinkles once every 1cm, the three-dimensional palate wrinkles are sequentially performed for 40 times, 40 two-dimensional palate wrinkles slice images are sequentially obtained after equidistant slicing, and are stored in a jpg format, and the 40 slice images together form a palate wrinkles slice sample set, as shown in fig. 5.
And thirdly, constructing a palate crease slice database. According to the unique characteristics of the palate wrinkle image, the method establishes a palate wrinkle tangent plane image database, acquires 5 palate wrinkle samples for 91 persons, performs the equidistant slicing operation of the step two on each palate wrinkle sample, and further acquires 40 palate wrinkle equidistant tangent plane curve images of each palate wrinkle sample, thereby obtaining 455 palate wrinkle samples and 18200 Zhang Ezhou tangent plane images of 91 persons. Wherein each person's 5 palate-wrinkled samples are three-dimensional palate-wrinkled images taken at different times, with the same light dark background and with a slight rotation.
And fourthly, performing image erosion and binarization on the palate crease slice. In order to eliminate irrelevant information in the image, restore useful real information, enhance the detectability of relevant information and simplify data to the greatest extent, thereby increasing the reliability of the extraction of the characteristics of the palate wrinkles, the segmentation of the palate wrinkles, the matching of the palate wrinkles and the identification.
And fifthly, extracting the cyclic spectrum characteristics of the palate crease line in the slice image. The method constructs a palate wrinkle characteristic vector by extracting the cyclic spectrum characteristic of a palate wrinkle tangent line, specifically, firstly, scanning the palate wrinkle tangent line image to obtain one-dimensional components in the horizontal direction and the vertical direction, and further, calculating spectrum correlation functions of the two one-dimensional components by adopting cumulative Fourier transformation. Assuming that the cyclic autocorrelation function of the discrete-time signal x (n) is
The spectrum correlation function is as follows:
in the formula (1), a is a cyclic frequency, and f is a spectral frequency. When a=0, the signal has a limited average power
The symmetry and discrete correlation periodicity of the spectral correlation function are shown in formulas (3) and (4):
in the formula (4), m and n are any integers.
The cyclic spectrum is a complex spectrum containing phase information that the power spectrum does not have, so that frequency information and phase information of the signal can be estimated by using a cyclic spectrum algorithm. Further, the spectral correlation function expression is shown in the formula (5):
in order to improve the calculation efficiency of cyclic spectrum estimation of a signal and reduce the calculation amount, a high-efficiency calculation algorithm is used for calculating a spectrum correlation function by accumulating Fourier transform, and the idea of the accumulated Fourier transform algorithm is to divide the total observation time length of the signal into a plurality of small segments, and then use a fast Fourier transform algorithm for carrying out cyclic frequency movement on each small segment. The spectral correlation function calculated by the cumulative fourier transform is:
in the formula (6), the amino acid sequence of the compound,
performing discrete digitization processing in the formula (6), the formula (6) becomes the formula (8):
in formula (8):
for ease of calculation, only the signal x (n) is considered to be a real sequence. When discretized, the sequence has a frequency of
The extraction flow of the cyclic spectrum features of the palate crease line is shown in fig. 7, and an example of a palate crease line image of a No. 37 palate crease sample in the sample library is shown in fig. 8. And carrying out cyclic spectrum estimation on the one-dimensional signal components in the horizontal direction and the vertical direction of the palate crease line to obtain a spectrum correlation function. On the contour diagram and the spectrum correlation function three-dimensional diagram, the partial characteristic extraction effect on the double frequency domains can be seen to be very good.
And step six, constructing a palate crease line circular spectrum characteristic vector. The cyclic spectral features of the palate crease line can be obtained using an accumulated fourier transform algorithm, how to construct the feature vectors for identification is an important issue. If the dual-frequency plan constructed by using the cumulative fourier transform algorithm has 64 cyclic spectrum estimation feature blocks, the energy and standard deviation of each small block are calculated, then feature vectors are constructed, and the dimension of the feature vectors is 128 (64 x 2) dimensions, in the method, each palate crease sample has 40 palate crease slice images, if the feature vectors are directly constructed according to the method, the dimension of the feature vectors can be high, which is not beneficial to classification calculation on one hand, and dimension disasters can be caused on the other hand. Therefore, in order to avoid dimension disasters and facilitate subsequent identification processing, the method adopts a 4×4 uniform block scheme to block the palate crease line circular spectrum characteristic diagram extracted in the step five, as shown in fig. 9, the block principle is as follows: the method comprises the steps of uniformly partitioning based on pixel positions, solving the mean value and the variance of each small block, and finally converting the mean value and the variance into one-dimensional row vectors in sequence, wherein a construction process schematic diagram of the palate crease tangent cycle frequency spectrum feature vector is shown in fig. 10. Compared with the traditional method for constructing the feature vector, the method based on the partitioning can realize the purpose of feature dimension reduction while constructing the feature vector, and the method can reduce the complexity of calculation and accelerate the speed of operation.
And seventhly, recognizing the three-dimensional palate wrinkles. The three-dimensional palate wrinkle recognition system adopted by the method comprises two components of a three-dimensional palate wrinkle image training process and a three-dimensional palate wrinkle image testing process, wherein each component comprises a three-dimensional palate wrinkle preprocessing module, a palate wrinkle tangent line circulation frequency spectrum characteristic extraction module, a palate wrinkle tangent line circulation frequency spectrum characteristic segmentation module and a palate wrinkle tangent line characteristic vector classification module, and a schematic diagram of the three-dimensional palate wrinkle recognition system is shown in fig. 11.
The specific method for three-dimensional palate crease identification comprises the following steps:
firstly, dividing a three-dimensional palate wrinkling digital image into a training sample and a test sample in a three-dimensional palate wrinkling preprocessing module, respectively sending the training sample and the test sample into a training process and a test process, respectively carrying out equidistant slicing treatment on the three-dimensional palate wrinkling training sample and the test sample in the two parts to obtain a three-dimensional palate wrinkling slice sample set, and carrying out corrosion and binarization treatment on the palate wrinkling slice image to construct a palate wrinkling slice database based on the three-dimensional palate wrinkling slice sample set;
secondly, respectively scanning the jaw wrinkle slice images in a training set and a test set in a jaw wrinkle tangent line circulation spectrum characteristic extraction module to acquire one-dimensional signal components of the jaw wrinkle tangent line in the horizontal direction and the vertical direction, respectively carrying out circulation spectrum analysis on the one-dimensional signal components in the two directions by using an accumulated Fourier transform algorithm, and extracting circulation spectrum characteristics of the two-dimensional signal components;
and finally, in the palate crease line circulation spectrum characteristic blocking module, a 4 multiplied by 4 uniform blocking scheme is adopted to block the palate crease line circulation spectrum characteristic graph extracted in the test set and the training set respectively, and the blocking principle is as follows: and uniformly partitioning based on pixel positions, solving the mean value and the variance of each small block, and finally converting the mean value and the variance into one-dimensional row vectors in sequence, so as to respectively construct feature vector dictionaries of a training set and a testing set.
Finally, a three-dimensional palate wrinkle sample label is added into a k nearest neighbor classifier in the palate wrinkle tangent line characteristic vector classification module, and the classifier outputs a classification result according to the provided information, wherein the classification result comprises a correct recognition rate, a correct error rate and a correct type label.
Specific cases:
the method comprises the steps of selecting samples from 11 th persons to 20 th persons in a database, establishing a small three-dimensional palate crease sample database, wherein each person has 5 three-dimensional samples, each sample has 40 palate crease tangent plane images obtained by equidistant slicing, selecting a palate crease tangent line construction sample database at the 40 th position on the assumption, selecting a palate crease tangent line construction sample database at the 39 th position, and so on to construct a palate crease tangent line database of 40 samples in total. In order to verify how many of the 40 palate crease lines of each palate crease sample can be successfully identified, the method selects the cyclic spectrum characteristics of the horizontal signal components of the palate crease lines for classification, and the results are shown in Table 1:
TABLE 1 identification rate influence of different Block schemes on palate crease lines at different positions
In Table 1, if the analysis is performed by taking the 11 th test sample as an example, 40 palate crease lines are completely identified correctly, the block scheme does not affect the identification rate, and the identification rate is kept at 100%; from analysis of samples 12 through 17, it can be seen that the number of identifiable palate crease lines increases as the number of segments increases. In fact, it can be found from the table that the number of correctly identifiable palate crease lines is more than 30 when the 4×8 block scheme and the 8×8 block scheme are adopted, which means that in some cases, palate crease lines at different positions can be used for identification.
In the whole three-dimensional palate wrinkling sample library, collecting palate wrinkles of 91 persons, 5 three-dimensional palate wrinkling samples of each person, selecting the first 4 three-dimensional palate wrinkles of each person as training samples, and 5 three-dimensional palate wrinkling samples of each person as test samples, wherein the identification conditions of the sample library are shown in table 2:
TABLE 2 statistics of different direction signal component recognition of palate crease line
In table 2, the number of recognition errors of the vertical and horizontal direction signal components and the corresponding recognition error class labels are given, and finally, the recognition of the two direction signal components achieves good recognition effect.
The foregoing description of the preferred embodiment of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.
Claims (6)
1. The three-dimensional palate crease identification method adopting the cyclic frequency spectrum is characterized by comprising the following specific steps of:
step one, collecting three-dimensional palate wrinkling information;
step two, equidistant slicing is carried out on the three-dimensional palate crease image;
step three, constructing a palate crease slice database;
step four, corroding and binarizing the palate crease slice image;
step five, extracting the cyclic spectrum characteristics of the palate crease line in the slice image;
step six, constructing a palate crease line circular spectrum feature vector;
step seven, recognizing the three-dimensional palate wrinkles and outputting a result;
in the first step, the position of a dental chair is adjusted to be in a position where the lower jaw of a volunteer is parallel to the ground and the upper jaw of the volunteer is perpendicular to the ground, the volunteer and the acquisition equipment are respectively disinfected, then the gun head of a scanning gun is stretched into the oral cavity of the volunteer, the scanning is carried out along teeth in a moving way, whether each surface of the palate crease is complete or not is checked, if the loophole exists or the scanning is incomplete, the supplementing scanning is carried out, and after the scanning is finished, the scanning data are imported into a computer to obtain three-dimensional palate crease data;
in the second step, the principle of equidistant slicing is as follows: intersecting the three-dimensional palate crease model with a horizontal slice plane to carry out slicing;
the specific operation of slicing is as follows: 1. measuring the length, width and height of the palate crease image by adopting image processing software, connecting the protruding points of the palate crease image into a cutting straight line from each of a plurality of teeth on the left and right sides of the mastoid of the incisors, then cutting from each of the five teeth on the left and right sides of the mastoid of the incisors by adopting a circular cutting line, adjusting the circular cutting line to be parallel to the cutting straight line, and carrying out equidistant slicing along the midpoint of the circular cutting line and the mastoid direction of the incisors; 2. and (3) starting equidistant slicing operation on the three-dimensional palate wrinkles at intervals of 1cm, sequentially performing 40 times of slicing operation, sequentially obtaining 40 two-dimensional palate wrinkles slice images after equidistant slicing, and storing the two-dimensional palate wrinkles slice images in jpg format, wherein the 40 slice images together form a palate wrinkles slice sample set.
2. The method according to claim 1, wherein in the third step, the jaw samples are collected for 91 persons, 5 jaw samples are collected for each person, and the equidistant slicing operation of the second step is performed for each jaw sample, so that 40 jaw equidistant tangent plane curve images of each jaw sample can be obtained, and thereby 455 jaw samples and 18200 Zhang Ezhou tangent plane images of 91 persons can be obtained.
3. A method of three-dimensional palate wrinkle identification using cyclic spectrum as defined in claim 2, wherein each of the 5 palate wrinkle samples is a three-dimensional palate wrinkle image obtained at different times, with the same light dark background and with rotation.
4. The method for recognizing three-dimensional palate wrinkles by using a cyclic spectrum according to claim 1, wherein in the fifth step, firstly, a palate wrinkle tangent image is scanned to obtain one-dimensional components in a horizontal direction and a vertical direction, and then a cumulative fourier transform is adopted to calculate spectrum correlation functions of the two one-dimensional components respectively;
setting the cyclic autocorrelation function of the discrete-time signal x (n) as R x a (k) The spectrum correlation function is:
in the formula (1), a is a cyclic frequency, and f is a spectral frequency; when a=0, the signal has a finite average power:
the symmetry and discrete correlation periodicity of the spectral correlation function are shown in formulas (3) and (4):
in the formula (4), m and n are any integers;
the cyclic spectrum is a complex spectrum, frequency information and phase information of signals are estimated by using a cyclic spectrum algorithm, and a spectrum correlation function expression is shown as a formula (5):
the spectral correlation function calculated by the cumulative fourier transform is:
in the formula (6), the amino acid sequence of the compound,
performing discrete digitization processing in the formula (6), the formula (6) becomes the formula (8):
in formula (8):
the signal x (n) is a real sequence whose frequency is, when discretized
5. The method for three-dimensional palate wrinkle identification with cyclic spectrum according to claim 1, wherein in step six, 4×4 uniform segmentation is performed on the palate wrinkle tangent cyclic spectrum feature map in step five, and the segmentation principle is as follows: and uniformly partitioning based on the pixel positions, solving the mean value and the variance of each small block, and sequentially converting the mean value and the variance into one-dimensional row vectors.
6. A method for three-dimensional palate wrinkle identification using cyclic spectrum as defined in claim 1, wherein in step seven, three-dimensional palate wrinkle identification comprises two parts of a three-dimensional palate wrinkle image training procedure and a three-dimensional palate wrinkle image testing procedure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911216905.1A CN110991339B (en) | 2019-12-02 | 2019-12-02 | Three-dimensional palate wrinkle identification method adopting cyclic frequency spectrum |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911216905.1A CN110991339B (en) | 2019-12-02 | 2019-12-02 | Three-dimensional palate wrinkle identification method adopting cyclic frequency spectrum |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110991339A CN110991339A (en) | 2020-04-10 |
| CN110991339B true CN110991339B (en) | 2023-04-28 |
Family
ID=70089354
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911216905.1A Active CN110991339B (en) | 2019-12-02 | 2019-12-02 | Three-dimensional palate wrinkle identification method adopting cyclic frequency spectrum |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110991339B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111753709A (en) * | 2020-06-19 | 2020-10-09 | 太原科技大学 | Fractional Fourier transform-based three-dimensional pucker-palate recognition method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103494653A (en) * | 2013-10-17 | 2014-01-08 | 福州大学 | Method for digitally measuring dental jaw growth and development laws |
| CN109363787A (en) * | 2018-12-04 | 2019-02-22 | 北京大学口腔医学院 | A kind of malocclusion denture rescues rear three-dimensional configuration high efficiency digitlization preferred method |
| CN208659317U (en) * | 2017-10-26 | 2019-03-29 | 山西医科大学口腔医院 | Palate wrinkle image collecting device |
| CN109948400A (en) * | 2017-12-20 | 2019-06-28 | 宁波盈芯信息科技有限公司 | It is a kind of to be able to carry out the smart phone and its recognition methods that face characteristic 3D is identified |
| CN110033032A (en) * | 2019-03-29 | 2019-07-19 | 中国科学院西安光学精密机械研究所 | A kind of histotomy classification method based on micro- high light spectrum image-forming technology |
Family Cites Families (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1801213A (en) * | 2005-09-30 | 2006-07-12 | 铁岭市公安局213研究所 | Method and apparatus for three-dimensional cranium body source identification |
| CN103078816A (en) * | 2013-01-18 | 2013-05-01 | 河海大学 | Interference resisting method for 2400 MHz (megahertz) frequency band |
| CN105096377B (en) * | 2014-05-14 | 2019-03-19 | 华为技术有限公司 | A kind of image processing method and device |
| ITUB20155817A1 (en) * | 2015-11-23 | 2017-05-23 | Torino Politecnico | PROCESS OF PROCESSING MEDICAL IMAGES OF A FACE FOR RECOGNITION OF FACIAL DISMORPHISMS |
| CN107169507A (en) * | 2017-01-06 | 2017-09-15 | 华南理工大学 | A kind of certificate photo exposure directions detection algorithm extracted based on grid search-engine |
| CN106875409B (en) * | 2017-03-24 | 2019-06-21 | 云南大学 | A kind of light-type incisional hernia sticking patch three-dimensional ultrasound pattern feature extracting method |
| CN107563983B (en) * | 2017-09-28 | 2020-09-01 | 上海联影医疗科技有限公司 | Image processing method and medical imaging device |
| CN109300486B (en) * | 2018-07-30 | 2021-06-25 | 四川大学 | PICGTFs and SSMC enhanced cleft palate speech pharynx fricative automatic identification method |
| CN109829885B (en) * | 2018-12-24 | 2022-07-22 | 广州柏视医疗科技有限公司 | Method for automatically identifying primary tumor of nasopharyngeal carcinoma based on deep semantic segmentation network |
| CN109802905B (en) * | 2018-12-27 | 2022-01-14 | 西安电子科技大学 | CNN convolutional neural network-based digital signal automatic modulation identification method |
| CN109801222B (en) * | 2019-01-25 | 2022-12-27 | 太原科技大学 | Method for correcting tilted images of pucker palate based on direction vectors |
| CN109784335B (en) * | 2019-01-25 | 2023-03-24 | 太原科技大学 | Method for calibrating boundary of region of interest of ruffle image based on least square fitting |
| CN110363803B (en) * | 2019-07-18 | 2021-08-27 | 光沦科技(深圳)有限公司 | Target detection method and system combining depth map slice and neural network |
-
2019
- 2019-12-02 CN CN201911216905.1A patent/CN110991339B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103494653A (en) * | 2013-10-17 | 2014-01-08 | 福州大学 | Method for digitally measuring dental jaw growth and development laws |
| CN208659317U (en) * | 2017-10-26 | 2019-03-29 | 山西医科大学口腔医院 | Palate wrinkle image collecting device |
| CN109948400A (en) * | 2017-12-20 | 2019-06-28 | 宁波盈芯信息科技有限公司 | It is a kind of to be able to carry out the smart phone and its recognition methods that face characteristic 3D is identified |
| CN109363787A (en) * | 2018-12-04 | 2019-02-22 | 北京大学口腔医学院 | A kind of malocclusion denture rescues rear three-dimensional configuration high efficiency digitlization preferred method |
| CN110033032A (en) * | 2019-03-29 | 2019-07-19 | 中国科学院西安光学精密机械研究所 | A kind of histotomy classification method based on micro- high light spectrum image-forming technology |
Non-Patent Citations (2)
| Title |
|---|
| 武秀萍等.应用数字图像识别系统研究正畸治疗前后腭皱形态变化.《中国法医学杂志》.2017,第32卷(第02期),第125-127、132页. * |
| 蔡志刚等.18三体腭裂小鼠上颌牙弓发育的 比较形态测量学研究.《口腔正畸学》.2000,第7卷(第1期),第34-37页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110991339A (en) | 2020-04-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109767841B (en) | Similar model retrieval method and device based on craniomaxillofacial three-dimensional morphological database | |
| CN110189324B (en) | Medical image processing method and processing device | |
| Syed et al. | Conversion of palatal rugae pattern to scanable Quick Response code in an Arabian population | |
| CN110991339B (en) | Three-dimensional palate wrinkle identification method adopting cyclic frequency spectrum | |
| Rai et al. | Using facial images for the diagnosis of genetic syndromes: a survey | |
| WO2015131091A1 (en) | Systems and methods for approximating the soft tissue profile of the skull of an unknown subject | |
| CN109377534A (en) | The non-linear oral cavity CT panoramic image synthesis method of automatic sampling thickness detection | |
| Tuan et al. | Shape prediction of nasal bones by digital 2D-photogrammetry of the nose based on convolution and back-propagation neural network | |
| Harastani et al. | Methamphetamine drug abuse and addiction: Effects on face asymmetry | |
| CN111753709A (en) | Fractional Fourier transform-based three-dimensional pucker-palate recognition method | |
| CN105913424B (en) | A kind of method and apparatus for inferring the age based on tooth | |
| Cattaneo et al. | Personal identification of cadavers and human remains | |
| Zhang et al. | Three-dimensional palatal rugae recognition based on cyclic spectral analysis | |
| CN115690045A (en) | Quantitative assessment method for bone increment before and after periodontitis treatment based on curved surface fault slice | |
| CN106447650B (en) | Human thyroglobulin method for measuring weight based on SPECT planar imaging | |
| KR101455198B1 (en) | Method, system and recordable medium for calculating quantitative assessment index of chest-wall deformities using ct image analysis | |
| CN111292289B (en) | CT lung tumor segmentation method, device, equipment and medium based on segmentation network | |
| Hori et al. | A novel identification method using perceptual degree of concordance of occlusal surfaces calculated by a Python program | |
| Karunya et al. | Human identification using dental biometrics | |
| Nuansanong et al. | Diagnosis and interpretation of dental X-ray in case of deciduous tooth extraction decision in children using active contour model and J48 tree | |
| Muad et al. | Tooth segmentation using dynamic programming-gradient inverse coefficient of variation | |
| Gerasimenko et al. | Results of processing CT scans of the jaw and preparing it for searching for the mandibular canal | |
| CN111222407B (en) | Palate wrinkle identification method adopting uniform slicing and inflection point characteristic extraction | |
| Henseler | Three-dimensional breast assessment by multiple stereophotogrammetry after breast reconstruction with latissimus dorsi flap | |
| Al-sherif | Novel Techniques for Automated Dental Identification |
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 |