CN105528795B - A kind of infrared face dividing method using annular shortest path - Google Patents
A kind of infrared face dividing method using annular shortest path Download PDFInfo
- Publication number
- CN105528795B CN105528795B CN201610090345.XA CN201610090345A CN105528795B CN 105528795 B CN105528795 B CN 105528795B CN 201610090345 A CN201610090345 A CN 201610090345A CN 105528795 B CN105528795 B CN 105528795B
- Authority
- CN
- China
- Prior art keywords
- image
- face
- reconstruction
- shortest path
- carried out
- 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
Classifications
-
- 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
- G06T2207/10048—Infrared image
-
- 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
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
-
- 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- 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
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
A kind of infrared face dividing method using annular shortest path, its implementation steps have：First, the infrared face image collected is pre-processed by morphological methods such as burn into expansions, obtains the image that feature is apparent, profile is more visible；2nd, binary conversion treatment is carried out using Otsu threshold to image, and operation is carried out out to it；3rd, the original image by pretreatment corresponding to pending human face region is sharpened；4th, data acquisition is carried out under the polar coordinates using geometric center as origin to combination picture.By above-mentioned steps, effective pretreatment is carried out to original infrared image, then by appropriate algorithm and optimization, has realized the Accurate Segmentation to infrared face.The present invention is widely used in all kinds of application systems based on image, has a vast market prospect and application value.
Description
(I) technical field
The invention relates to an infrared human face segmentation method utilizing a shortest annular path, belongs to the field of digital image processing, and mainly relates to the shortest annular path and an image segmentation technology. The method has wide application prospect in various image-based application systems.
(II) background of the invention
Image segmentation is a technique and process for dividing an image into specific regions with unique properties and extracting an object of interest, and is a key step from image processing to image analysis. With the intensive research of image segmentation technology, the application of the image segmentation technology is becoming more and more extensive. The image segmentation is required to be carried out on an image target, and the technology is widely applied to the fields of traffic, medicine, remote sensing, communication, military, industrial automation and the like. Researchers have proposed many methods for image segmentation. Threshold-based algorithms (see the literature: otsu. A Threshold Selection Method based on Gray histogram, american institute of Electrical and electronics Engineers System, proc. And treaty on control, 1979,9, 62-66. (Otsu N.A Threshold Selection Method from Gray-Scale Histograms [ J ]. IEEE Trans on Smc,1979, 62-66.)) use the Gray histogram information of images to obtain thresholds for segmentation, and the class variance Method and the maximum entropy Method are two of the most common Threshold detection methods. The threshold-based algorithm has the advantages of simple implementation and high operation speed, and has the defect of difficulty in processing the condition containing a plurality of foreground objects. Edge-based segmentation (see Canny, a Computational method for Edge Detection, the institute of Electrical and electronics Engineers (American society of Electrical and electronics Engineers) model Analysis and Machine Intelligent journal, 1986, PAMI-8 (6): 679-698 (Canny J,. A Computational Approach to Edge Detection [ J ]. Pattern Analysis & Machine Intelligent Transactions on,1986, PAMI-8 (6): 679-698)), edge-based segmentation is segmentation by detecting edges of different regions. By edge is meant the set of pixels around which there is a step change in the grey level of the pixels. It exists between the object and the background and is the most important feature on which image segmentation depends. The Region Growing method (see Adams et al, region Growing method, american society of Electrical and electronics Engineers, pattern Analysis and Machine Intelligent Association, 1994,16 (6): 641-647 (Adams R, bischof L. Selected Region Growing [ J ]. IEEE Transactions on Pattern Analysis & Machine understanding, 1994,16 (6): 641-647)) divides a Region by using local spatial information, groups pixels having similar characteristics to constitute a Region, and mainly includes the Region Growing method and the split-merge method. Clustering-based Algorithms (see, e.g., kann et al., effective application of Fuzzy C-Means Clustering algorithm, american society of Electrical and electronic Engineers, japan, 1986, PAMI-8 (2): 248-255 (Cannon R L, dave J V, bezdek J C. Efficient implantation of the Fuzzy C-Means Clustering Algorithms [ J ]. IEEE Transactions on Pattern Analysis & Machine integration, 1986, PAMI-8 (2): 248-255)) Clustering image segmentation as a Clustering problem are widely used in image segmentation, such as K-Means algorithm, FCM (Fuzzy C-Means) Fuzzy algorithm, ISODATA algorithm, etc. Wavelet analysis based algorithms (see the literature: anthony et al, image coding based on wavelet transform, journal of Image Processing by the Electrical and electronic Engineers in America, 1992,1 (2): 205-220 (Antonini M, barlaud M, mathieu P. Image coding using wavelet transform [ J ]. IEEE Transactions on Image Processing,1992,1 (2): 205-220)) have good time-frequency localization characteristics, scale change characteristics and orientation characteristics, and have good effects in the aspects of Image Processing, computer vision, texture analysis and the like. Neural network theory and technology are introduced into the image space clustering segmentation field by using neural network-based algorithm (see the literature: yang, color image text positioning based on neural network, and Pattern Recognition statement, 2001,22 (14): 1503-1515. (Jung K. Neural network-based text positioning in color images [ J ]. Pattern Recognition Letters,2001,22 (14): 1503-1515.)), so that the limitation of the using condition of the traditional clustering method is broken, and a foundation is laid for constructing various new clustering methods. An algorithm based on the shortest circular path (see document: sun, et al, application of shortest circular path in image processing, pattern Recognition,2003,36 (36): 709-719 (Sun C, pixel s. Circular short path in images [ J ]. Pattern Recognition,2003,36 (36))) can be used to solve the segmentation problem of the circle-like region, and the image edge is found by using the shortest path algorithm through the conversion of polar coordinates and a rectangular coordinate system.
The infrared face segmentation is an image segmentation technology for processing infrared band face imaging, and the acquired infrared face imaging in different bands is subjected to image segmentation processing to obtain an accurate face region, so that the face region is preprocessed for next detection and identification. (see literature: philips et al, a New thermal Infrared Face Segmentation Method for Pose transformation, computer Science series, 2013, 7887.
The morphological operation can effectively segment the foreground and the background, an area with common characteristics is extracted, a more accurate face range can be obtained through multiple times of morphological processing and threshold selection, the annular shortest path algorithm can effectively extract the image edge similar to annular distribution, and the method is suitable for fine segmentation of the face. Therefore, the special advantages of morphological processing and the shortest circular path are reasonably utilized, and a good effect can be brought to a specific human face segmentation scene. In order to more effectively extract the infrared face area, the invention provides an infrared face segmentation method utilizing the shortest annular path.
Disclosure of the invention
1. The purpose is as follows: the infrared face detection is a new technology which is developed rapidly and applied more and more widely, but the key link of the infrared face detection still has a great blank and challenge at present. The traditional image segmentation algorithm cannot be well adapted to a unique scene under an infrared condition, so that the infrared face cannot be presented most accurately.
In order to solve the problems and make up for the defects of the traditional method, the invention provides the infrared face segmentation method utilizing the annular shortest path, which effectively segments the foreground and the background through morphological processing to obtain a more accurate face connected region, effectively extracts the face region of the face through the annular shortest path to further realize accurate infrared face segmentation, and avoids excessive segmentation of the face while realizing accurate extraction of the image foreground. The annular shortest path algorithm can effectively extract more detailed face edges and can effectively solve special problems existing in the infrared image.
2. The technical scheme is as follows: in order to realize the purpose, the technical scheme of the invention is that firstly, the collected infrared portrait is preprocessed by opening, closing, corroding, expanding and other morphological methods to obtain an image with obvious characteristics and clear outline; then, carrying out binarization processing on the image, carrying out opening operation on the image, taking out the largest connected region, obtaining a region to be processed containing the face, and calculating the geometric center of the region; secondly, sharpening and gradient calculating are carried out on the preprocessed original image corresponding to the face area to be processed, and then the two operation results are fused into a composite image; and finally, acquiring data of the composite image under a polar coordinate with a geometric center as an original point, negating and normalizing the data, expanding the data into a rectangular coordinate system, establishing a penalty function on an annular track, solving the shortest path of the composite image under the condition of head-tail adjacent constraint, converting the shortest path finally obtained into the original polar coordinate system, and enabling the presented closed annular track to be the obtained accurate contour of the human face, thereby realizing the accurate segmentation of the infrared human face.
The invention relates to an infrared face segmentation method by using an annular shortest path, which is characterized by comprising the following steps of: the method comprises the following steps:
the original face image is recorded as f:
the method comprises the following steps: the acquired infrared human face image is preprocessed by morphological methods such as corrosion, expansion and the like to obtain an image with obvious characteristics and clear outline:
f reconstruction ＝f Expansion of ∷f
Wherein, f is an original face image (called as an original image for short); b is a structural element; the: is the morphological reconstruction (left is the marker of the morphological reconstruction and right is the template of the morphological reconstruction);andmorphological dilation and erosion operators, respectively; step two: and (5) carrying out binarization processing on the image by adopting an Otsu threshold value, and carrying out opening operation on the image.
f Binarization method ＝f Reconstruction ○(f Reconstruction (intensity＞graythresh(f Reconstruction )))
Where O is the morphological open operator.
And taking out the largest communication area, namely the area to be processed containing the face:
f face part ＝f Reconstruction (f Connected region ＞0)
And calculates the geometric center of the region:
wherein graythresh is the Otsu threshold; f. of Face part Is the portion of the processed image in the region of the face to be processed, x Center of a ship And y Center of a ship Respectively are the x-direction central point coordinates and the y-direction central point coordinates (the original points of polar coordinate expansion) of the obtained human face area to be processed.
Step three: sharpening the preprocessed original image corresponding to the face area to be processed (respectively solving secondary partial derivatives in the x direction and the y direction):
and then, threshold segmentation is carried out on the original image:
f threshold value ＝f Face part (intensity>graythresh(f Face part ))
Their gradients are then recalculated:
and fusing the two operation results into a composite image:
f fusion of ＝a 1 ·f Gradient 1 +a 2 ·f Gradient 2
Wherein a is 1 And a 2 Is a corresponding weighting coefficient, a 1 ＝0.95,a 2 ＝1。
Step four: data acquisition is carried out on the composite image under polar coordinates with the geometric center as an origin, and f Polar coordinates I.e. in step three Fusion ：
C 0 (i,p i )＝f Polar coordinates (r,θ)
And then, negating and normalizing the data, and establishing a penalty function on the circular track:
the method is expanded into a rectangular coordinate system, and the shortest path solution under the condition of head-tail adjacent constraint is carried out on the rectangular coordinate system by using dynamic programming,
and the CSP is the shortest path finally obtained, and the shortest path is expanded into a rectangular coordinate system to form the face contour.
Through the steps, the original infrared image is effectively preprocessed, and accurate segmentation of the infrared face is achieved through proper algorithm and optimization.
3. The advantages and the effects are as follows: the infrared human face image has larger difference with the human face image under the visible light, and a specific method needs to be provided for processing.
Although many image segmentation methods can effectively segment faces, they are not satisfactory for processing infrared faces. The results of this experiment fully demonstrate the effectiveness of the present invention. Moreover, the image for experiment is collected from various different postures of different experimental objects in a real scene, and the method has good experimental effect, so that the method fully shows that the method can be widely applied to application systems, and has wide market prospect and application value.
(IV) description of the drawings
Fig. 1 is a flow chart of an infrared face segmentation method using a shortest circular path according to the present invention.
In the figure, the symbol f is an original face image
(V) detailed description of the preferred embodiments
For better understanding of the technical solutions of the present invention, the following further describes embodiments of the present invention with reference to the accompanying drawings. The flow chart of the invention is shown in figure 1, the invention relates to an infrared human face segmentation method using the shortest circular path, which comprises the following concrete implementation steps:
firstly, preprocessing an acquired infrared human face image by morphological methods such as opening, closing, corrosion, expansion and the like to obtain an image with obvious characteristics and clear outline; then, carrying out binarization processing on the image, then carrying out opening operation on the image, taking out the largest connected region, namely the region to be processed containing the human face, and calculating the geometric center of the region; secondly, sharpening and gradient calculating are carried out on the preprocessed original image corresponding to the face area to be processed, and then the two operation results are fused into a composite image; and finally, acquiring data of the composite image under a polar coordinate with the geometric center as an origin, negating and normalizing the data, expanding the data into a rectangular coordinate system, establishing a corresponding penalty function, solving the shortest path under the condition of head-tail adjacent constraint, converting the shortest path finally obtained into the original polar coordinate system, and converting the presented closed annular track into the obtained accurate contour of the face, thereby realizing the accurate segmentation of the infrared face. The invention relates to an infrared face segmentation method using a shortest annular path, which comprises the following specific steps:
the original face image is noted as f:
the method comprises the following steps: preprocessing the collected infrared human face image by morphological methods such as corrosion, expansion and the like to obtain an image with obvious characteristics and clear outline:
f reconstruction ＝f Expansion of ∷f
Wherein, f is an original face image; b is a structural element; the: is the morphological reconstruction (marker for morphological reconstruction on the left, template for morphological reconstruction on the right);andmorphological dilation and erosion operators, respectively; through a series of morphological processing, the interference information in the image can be removed, a face image with better consistency is obtained, and the next step of area selection is facilitated.
Step two: and (5) carrying out binarization processing on the image by adopting an Otsu threshold value, and carrying out opening operation on the image.
f Binarization method ＝f Reconstruction ○(f Reconstruction (intensity＞graythresh(f Reconstruction )))
Where O is the morphological open operator.
And taking out the largest communication area, namely the area to be processed containing the face:
f face part ＝f Reconstruction (f Communicating region ＞0)
Respectively carrying out integral projection on the binarized image in the X direction and the Y direction, and recording the coordinates corresponding to the maximum value of the two projection curves as the central point (X) of the face Center of a ship And y Center (C) ) This is used as the origin of the polar coordinate conversion in the following operation:
wherein graythresh is the Otsu threshold; f. of Face part Is the part of the processed image in the region of the face to be processed, x Center of a ship And y Center of a ship Respectively are the x-direction central point coordinates and the y-direction central point coordinates (the original points of polar coordinate expansion) of the obtained human face area to be processed.
Step three: sharpening the preprocessed original image corresponding to the face area to be processed (respectively solving secondary partial derivatives in the x direction and the y direction):
and then, threshold segmentation is carried out on the original image:
f threshold value ＝f Face part (intensity>graythresh(f Face part ))
Their gradients are then recalculated:
and fusing the two operation results into a composite image:
f fusion of ＝a 1 ·f Gradient 1 +a 2 ·f Gradient 2
Wherein a is 1 And a 2 Are corresponding weighting coefficients, a 1 ＝0.95,a 2 ＝1。
Step four: data acquisition is performed on the composite image in polar coordinates with the geometric center as the origin, f Polar coordinates I.e. in step three Fusion of ：
C 0 (i,p i )＝f Polar coordinates (r,θ)
And then, negating and normalizing the data, and establishing a penalty function on the circular track:
the method is expanded into a rectangular coordinate system, and the shortest path under the condition of head-tail adjacent constraint is solved by using dynamic programming,
and the CSP is the shortest path which is finally solved, and the CSP is expanded into a rectangular coordinate system to obtain the face contour.
The infrared human face image has larger difference with the human face image under the visible light, a specific method needs to be provided for processing, and aiming at different characteristics, a plurality of processing methods are adopted, so that necessary information is enhanced, and a good operation environment is provided for the later processing.
The experimental data set collects 400 infrared images of 40 persons with different postures, and the images are shot by using a medium-long wave infrared camera at a shooting distance of 1.5 m. The invention uses morphological operation to preprocess the image, effectively strengthens the edge information in the original image, further strengthens the edge in the original image by two different processing methods and gradient extraction, and finally uses the annular shortest path to extract the edge similar to a circle, thereby obtaining the accurate face contour.
Although many image segmentation methods can effectively segment faces, they are not satisfactory for processing infrared faces. The results of this experiment fully demonstrate the effectiveness of the present invention. Moreover, the image for experiment is collected from various different postures of different experimental objects in a real scene, and the method has good experimental effect, so that the method fully shows that the method can be widely applied to application systems, and has wide market prospect and application value.
Claims (1)
1. An infrared human face segmentation method using a ring shortest path is characterized in that: the method comprises the following steps:
the method comprises the following steps: preprocessing the collected infrared human face image by a corrosion and expansion morphological method to obtain an image with obvious characteristics and clear outline:
f reconstruction ＝f Expansion of Chinese character' Ji f
Wherein f is an original face image; b is a structural element; the: is the morphological reconstruction;andmorphological dilation and erosion operators, respectively;
step two: carrying out binarization processing on the image by adopting an Otsu threshold value, and carrying out opening operation on the image:
f binarization method ＝f Reconstruction ○(f Reconstruction (intensity＞graythresh(f Reconstruction )))
Wherein O is the morphological opening operator;
and taking out the largest connected region, namely the region to be processed containing the face:
f face part ＝f Reconstruction (f Connected region ＞0)
And calculates the geometric center of the region:
wherein graythresh is the Otsu threshold; f. of Face part Is the portion of the processed image in the region of the face to be processed, x Center (C) And y Center of a ship Respectively unfolding the x and y coordinates of the central point of the obtained face area to be processed, namely the polar coordinates;
step three: sharpening the preprocessed original image corresponding to the face area to be processed, namely respectively solving secondary partial derivatives in the x direction and the y direction:
and then, threshold segmentation is carried out on the original image:
f threshold value ＝f Face part (intensity>graythresh(f Face part ))
Their gradients are then recalculated:
and fusing the two operation results into a composite image:
f fusion of ＝a 1 ·f Gradient 1 +a 2 ·f Gradient 2
Wherein a is 1 And a 2 Are corresponding weighting coefficients, a 1 ＝0.95,a 2 ＝1；
Step four: data acquisition is carried out on the composite image under polar coordinates with the geometric center as an origin, and f Polar coordinates I.e. in step three Fusion ：
C 0 (i,p i )＝f Polar coordinates (r,θ)
And then, negating and normalizing the data, and establishing a penalty function on the circular track:
the method is expanded into a rectangular coordinate system, and the shortest path solution under the condition of head-tail adjacent constraint is carried out on the rectangular coordinate system by using dynamic programming,
and the CSP is the shortest path finally obtained, and the shortest path is converted into the original polar coordinate system to obtain the final face contour.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610090345.XA CN105528795B (en) | 2016-02-18 | 2016-02-18 | A kind of infrared face dividing method using annular shortest path |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610090345.XA CN105528795B (en) | 2016-02-18 | 2016-02-18 | A kind of infrared face dividing method using annular shortest path |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105528795A CN105528795A (en) | 2016-04-27 |
| CN105528795B true CN105528795B (en) | 2018-06-01 |
Family
ID=55770998
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610090345.XA Active CN105528795B (en) | 2016-02-18 | 2016-02-18 | A kind of infrared face dividing method using annular shortest path |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105528795B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106296949B (en) * | 2016-09-13 | 2018-11-27 | 成都创慧科达科技有限公司 | A kind of access control system and method based on H264 video decoding techniques |
| CN106652112A (en) * | 2016-09-24 | 2017-05-10 | 成都创慧科达科技有限公司 | Intelligent multifunctional door lock, control system and control method |
| CN111839519B (en) * | 2020-05-26 | 2021-05-18 | 合肥工业大学 | Non-contact respiratory frequency monitoring method and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101944232A (en) * | 2010-09-02 | 2011-01-12 | 北京航空航天大学 | Precise segmentation method of overlapped cells by using shortest path |
| CN103020948A (en) * | 2011-09-28 | 2013-04-03 | 中国航天科工集团第二研究院二○七所 | Night image characteristic extraction method in intelligent vehicle-mounted anti-collision pre-warning system |
| CN104766321A (en) * | 2015-04-03 | 2015-07-08 | 北京航空航天大学 | Infrared pedestrian image accurate segmentation method utilizing shortest annular path |
-
2016
- 2016-02-18 CN CN201610090345.XA patent/CN105528795B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101944232A (en) * | 2010-09-02 | 2011-01-12 | 北京航空航天大学 | Precise segmentation method of overlapped cells by using shortest path |
| CN103020948A (en) * | 2011-09-28 | 2013-04-03 | 中国航天科工集团第二研究院二○七所 | Night image characteristic extraction method in intelligent vehicle-mounted anti-collision pre-warning system |
| CN104766321A (en) * | 2015-04-03 | 2015-07-08 | 北京航空航天大学 | Infrared pedestrian image accurate segmentation method utilizing shortest annular path |
Non-Patent Citations (1)
| Title |
|---|
| Circular shortest path in images;Changming Sun et al.;《Pattern Recognition》;20030330;第36卷;第711-721页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105528795A (en) | 2016-04-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107578035B (en) | Human body contour extraction method based on super-pixel-multi-color space | |
| CN110084156B (en) | Gait feature extraction method and pedestrian identity recognition method based on gait features | |
| CN106097335B (en) | Alimentary canal lesion image identification system and recognition methods | |
| CN108319973B (en) | Detection method for citrus fruits on tree | |
| Guan et al. | Accurate segmentation of partially overlapping cervical cells based on dynamic sparse contour searching and GVF snake model | |
| Konwar et al. | An American sign language detection system using HSV color model and edge detection | |
| CN110738676A (en) | GrabCT automatic segmentation algorithm combined with RGBD data | |
| CN112819094A (en) | Target detection and identification method based on structural similarity measurement | |
| CN110728302A (en) | Method for identifying color textile fabric tissue based on HSV (hue, saturation, value) and Lab (Lab) color spaces | |
| CN104484652A (en) | Method for fingerprint recognition | |
| CN105528795B (en) | A kind of infrared face dividing method using annular shortest path | |
| CN102147867A (en) | Method for identifying traditional Chinese painting images and calligraphy images based on subject | |
| CN111080574A (en) | Fabric defect detection method based on information entropy and visual attention mechanism | |
| CN111080696A (en) | Underwater sea cucumber identification and positioning method based on computer vision | |
| Zhang et al. | Multiple Saliency Features Based Automatic Road Extraction from High‐Resolution Multispectral Satellite Images | |
| Das et al. | Human skin region segmentation based on chrominance component using modified watershed algorithm | |
| CN111161276B (en) | Iris normalized image forming method | |
| CN108710883A (en) | A kind of complete conspicuousness object detecting method using contour detecting | |
| Senthilkani et al. | Overlap wavelet transform for image segmentation | |
| Wang et al. | Lane detection algorithm based on density clustering and RANSAC | |
| Chen | Superpixel segmentation based on Delaunay Triangulation | |
| Samet et al. | An implementation of automatic contour line extraction from scanned digital topographic maps | |
| Li et al. | Automatic segmentation of color lip images based on morphological filter | |
| CN114926635B (en) | Target segmentation method in multi-focus image combined with deep learning method | |
| CN115578629A (en) | Highway guardrail extraction method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |