CN111080663B - Bionic contour detection method based on dynamic receptive field - Google Patents
Bionic contour detection method based on dynamic receptive field Download PDFInfo
- Publication number
- CN111080663B CN111080663B CN201911392146.4A CN201911392146A CN111080663B CN 111080663 B CN111080663 B CN 111080663B CN 201911392146 A CN201911392146 A CN 201911392146A CN 111080663 B CN111080663 B CN 111080663B
- Authority
- CN
- China
- Prior art keywords
- pixel point
- receptive field
- value
- response value
- standard deviation
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 28
- 239000011664 nicotinic acid Substances 0.000 title claims abstract description 12
- 238000000034 method Methods 0.000 claims abstract description 11
- 230000003042 antagnostic effect Effects 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 6
- 230000001629 suppression Effects 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 5
- 230000007547 defect Effects 0.000 abstract description 2
- 230000000007 visual effect Effects 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000026676 system process Effects 0.000 description 1
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
- G06T7/13—Edge detection
-
- 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
Abstract
The invention aims to provide a bionic contour detection method based on dynamic receptive fields, which comprises the following steps: A. inputting a to-be-detected image subjected to gray processing, and calculating a spatial standard deviation of each pixel point; B. calculating the classical receptive field optimal response value of each pixel point by combining the spatial standard deviation of each pixel point; C. calculating the final contour response value of each pixel point; D. and calculating the final contour value of each pixel point. The method overcomes the defects of the prior art and has the characteristics of good bionic effect and high outline detection rate.
Description
Technical Field
The invention belongs to the field of image contour detection, and particularly relates to a bionic contour detection method based on a dynamic receptive field.
Background
Contour detection is an important basis for technologies such as target detection, shape analysis, target recognition, target tracking, and the like. In a visual system, the dynamics are all the way through the process of contour extraction. Because of the difference of local image information, the size of the receptive field is dynamically changed when the visual system processes information, but many contour detection models inspired by biology at present simply simulate a part of physiological characteristics in the visual system, and in the contour extraction process, some visual dynamic characteristics cannot be well simulated, so that the problems of the lack of contour information and the enhancement of texture information are caused to a certain extent, and the integrity of the target contour cannot be well ensured.
Disclosure of Invention
The invention aims to provide a bionic type contour detection method based on a dynamic receptive field, which overcomes the defects of the prior art and has the characteristics of good bionic effect and high contour detection rate.
The technical scheme of the invention is as follows: a bionic contour detection method based on dynamic receptive fields comprises the following steps:
A. inputting a to-be-detected image subjected to gray processing, and calculating a spatial standard deviation of each pixel point;
B. calculating the classical receptive field optimal response value of each pixel point by combining the spatial standard deviation of each pixel point;
C. calculating the final contour response value of each pixel point;
D. and calculating the final contour value of each pixel point.
Preferably, the specific steps are as follows:
A. inputting an image to be detected after gray processing, presetting a local area and the radius of the local area, wherein the local area is square, and the side length of the local area is twice of the radius of the local area; for each pixel point, calculating the standard deviation of the gray value of each pixel point in a local area taking the pixel point as the center, and calculating the obtained standard deviation to obtain the spatial standard deviation of each pixel point;
B. presetting a two-dimensional Gaussian first-order partial derivative function containing a plurality of direction parameters, and filtering the gray value of each pixel point by adopting the two-dimensional Gaussian first-order partial derivative function, wherein the standard deviation in the two-dimensional Gaussian first-order partial derivative function is the spatial standard deviation corresponding to each pixel point, so as to obtain the classical receptive field initial response value of each direction parameter of each pixel point; for each pixel point, respectively taking the maximum value of the classical receptive field initial response value of each directional parameter of the pixel point, and taking the maximum value as the optimal response value of the classical receptive field of the pixel point;
C. presetting a normalized Gaussian difference function and non-classical receptive field antagonistic strength, and filtering the optimal response value of the classical receptive field of each pixel point by adopting the normalized Gaussian difference function to obtain the non-classical receptive field response value of each pixel point; for each pixel point, subtracting the product of the non-classical receptive field response value and the non-classical receptive field antagonistic strength from the classical receptive field optimal response value of the pixel point to obtain the final contour response value of each pixel point;
D. and for each pixel point, carrying out non-maximum suppression and double-threshold processing on the final contour response value contour value of each pixel point to obtain the final contour value of each pixel point.
Preferably, the step a is as follows:
the standard deviation
The spatial standard deviation
Wherein R isxyDenotes a local region, R denotes a radius of the local region, and the local region RxyHas a side length of (2r +1),
is the gray value of the ith pixel in the local region, N is the number of pixel points in the local region, muRThe gray value mean value of each pixel point in the local area is obtained.
Preferably, the step B is as follows:
the two-dimensional Gaussian first-order partial derivative function
classical receptive field initial response value E (x, y, theta) of each pixel point in each direction parameteri)＝I(x,y)*G(x,y,θi) (4)；
Wherein I (x, y) is the gray value of each pixel point;
the optimal response value E (x, y) of the classical receptive field of each pixel point is max (E (x, y, theta)i)|i＝1,2,…Nθ) (5)。
Preferably, the step C is specifically as follows:
the non-classical receptive field response value Inh (x, y) ═ E (x, y) × W (x, y) (6) of each pixel point;
normalized difference function of gaussians
Wherein the content of the first and second substances,
wherein | · | purple1Is L1Norm, h (x) max (0, x);
a final contour response value O (x, y) ═ E (x, y) - β · Inh (x, y) (7) of each pixel point;
wherein β is the non-classical receptor antagonistic strength.
The invention adopts the characteristic of local standard deviation to simulate the size change of the visual receptive field, thereby simulating the change of the complexity of a simulation image to the maximum extent for contour detection, realizing the simulation of the dynamic change of the size of the receptive field, and further adopting the corresponding spatial receptive field for each pixel point when calculating the classical receptive field response and the non-classical receptive field response, thereby improving the simulation degree, the identification accuracy and the robustness of the contour detection.
Drawings
FIG. 1 is a flow chart of a contour detection method of the present invention;
fig. 2 is a comparison graph of the detection effect of the method of example 1 and the detection effect of the contour detection model of document 1.
Detailed Description
The present invention will be described in detail with reference to the accompanying drawings and examples.
Example 1
The bionic contour detection method based on the dynamic receptive field provided by the embodiment comprises the following steps:
A. inputting an image to be detected after gray processing, presetting a local area and the radius of the local area, wherein the local area is square, and the side length of the local area is twice of the radius of the local area; for each pixel point, calculating the standard deviation of the gray value of each pixel point in a local area taking the pixel point as the center, and calculating the obtained standard deviation to obtain the spatial standard deviation of each pixel point;
the step A is specifically as follows:
the standard deviation
The spatial standard deviation
Wherein R isxyDenotes a local region, R denotes a radius of the local region, and the local region RxyHas a side length of (2r +1),
is the gray value of the ith pixel in the local region, N is the number of pixel points in the local region, muRThe gray value mean value of each pixel point in the local area is obtained;
B. presetting a two-dimensional Gaussian first-order partial derivative function containing a plurality of direction parameters, and filtering the gray value of each pixel point by adopting the two-dimensional Gaussian first-order partial derivative function, wherein the standard deviation in the two-dimensional Gaussian first-order partial derivative function is the spatial standard deviation corresponding to each pixel point, so as to obtain the classical receptive field initial response value of each direction parameter of each pixel point; for each pixel point, respectively taking the maximum value of the classical receptive field initial response value of each directional parameter of the pixel point, and taking the maximum value as the optimal response value of the classical receptive field of the pixel point;
the step B is specifically as follows:
the two-dimensional Gaussian first-order partial derivative function
classical receptive field initial response value E (x, y, theta) of each pixel point in each direction parameteri)＝I(x,y)*G(x,y,θi) (4)；
Wherein I (x, y) is the gray value of each pixel point;
the optimal response value E (x, y) of the classical receptive field of each pixel point is max (E (x, y, theta)i)|i＝1,2,…Nθ) (5)；
C. Presetting a normalized Gaussian difference function and non-classical receptive field antagonistic strength, and filtering the optimal response value of the classical receptive field of each pixel point by adopting the normalized Gaussian difference function to obtain the non-classical receptive field response value of each pixel point; for each pixel point, subtracting the product of the non-classical receptive field response value and the non-classical receptive field antagonistic strength from the classical receptive field optimal response value of the pixel point to obtain the final contour response value of each pixel point;
the step C is specifically as follows:
the non-classical receptive field response value Inh (x, y) ═ E (x, y) × W (x, y) (6) of each pixel point;
normalized difference function of gaussians
Wherein the content of the first and second substances,
wherein | · | purple1Is L1Norm, h (x) max (0, x);
a final contour response value O (x, y) ═ E (x, y) - β · Inh (x, y) (7) of each pixel point;
wherein β is the non-classical receptor field antagonistic strength;
D. and for each pixel point, carrying out non-maximum suppression and double-threshold processing on the final contour response value contour value of each pixel point to obtain the final contour value of each pixel point.
The following compares the effectiveness of the contour detection method of the present embodiment with the contour detection method provided in document 1, where document 1 is as follows:
document 1: yang K F, Li C Y, Li Y J, Multi-feature-based failure detection in natural images [ J ]. IEEE Transactions on image processing,2014,23(12): 5020-5032;
to ensure the effectiveness of the comparison, the same non-maximum suppression and double-threshold processing as in document 1 are used for the final contour integration for this embodiment, wherein two thresholds t are includedh,tlIs set to tl＝0.5thCalculated from a threshold quantile p;
wherein the performance evaluation index F employs the following criteria given in document 1:
wherein P represents the accuracy, R represents the recall rate, the value of the performance evaluation index F is between [0,1], the closer to 1, the better the effect of the contour detection is represented, and in addition, the definition tolerance is as follows: all detected within 5 x 5 neighbourhoods are counted as correct detections.
Selecting three random natural images of a Berkeley segmentation data set (BSDS300), and respectively adopting the scheme of embodiment 1 and the scheme of document 1 to detect, wherein the corresponding real profile and the optimal profile detected by the method of document 1 are shown in FIG. 2; in the optimal profile graph detected by the method in document 1, the number at the upper right corner in the optimal profile graph detected by the method in embodiment 1 is the value of the corresponding performance evaluation index F, and table 1 is the parameter value selected in embodiment 1;
table 1 example 1 parameter set table
As can be seen from fig. 2, the contour detection result of the embodiment 1 is superior to that of the document 1.
Claims (3)
1. A bionic contour detection method based on dynamic receptive fields is characterized by comprising the following steps:
A. inputting a to-be-detected image subjected to gray processing, and calculating a spatial standard deviation of each pixel point; the method specifically comprises the following steps: inputting an image to be detected after gray processing, presetting a local area and the radius of the local area, wherein the local area is square, and the side length of the local area is twice of the radius of the local area; for each pixel point, calculating the standard deviation of the gray value of each pixel point in a local area taking the pixel point as the center, and calculating the obtained standard deviation to obtain the spatial standard deviation of each pixel point;
the standard deviation
The spatial standard deviation
Wherein R isxyDenotes a local region, R denotes a radius of the local region, and the local region RxyHas a side length of (2r +1),
is the gray value of the ith pixel in the local region, N is the number of pixel points in the local region, muRThe gray value mean value of each pixel point in the local area is obtained;
B. calculating the classical receptive field optimal response value of each pixel point by combining the spatial standard deviation of each pixel point; the method specifically comprises the following steps: presetting a two-dimensional Gaussian first-order partial derivative function containing a plurality of direction parameters, and filtering the gray value of each pixel point by adopting the two-dimensional Gaussian first-order partial derivative function, wherein the standard deviation in the two-dimensional Gaussian first-order partial derivative function is the spatial standard deviation corresponding to each pixel point, so as to obtain the classical receptive field initial response value of each direction parameter of each pixel point; for each pixel point, respectively taking the maximum value of the classical receptive field initial response value of each directional parameter of the pixel point, and taking the maximum value as the optimal response value of the classical receptive field of the pixel point;
C. calculating the final contour response value of each pixel point; the method specifically comprises the following steps: presetting a normalized Gaussian difference function and non-classical receptive field antagonistic strength, and filtering the optimal response value of the classical receptive field of each pixel point by adopting the normalized Gaussian difference function to obtain the non-classical receptive field response value of each pixel point; for each pixel point, subtracting the product of the non-classical receptive field response value and the non-classical receptive field antagonistic strength from the classical receptive field optimal response value of the pixel point to obtain the final contour response value of each pixel point;
D. calculating the final contour value of each pixel point; the method specifically comprises the following steps: and for each pixel point, carrying out non-maximum suppression and double-threshold processing on the final contour response value contour value of each pixel point to obtain the final contour value of each pixel point.
2. The method of claim 1, wherein the bionic profile detection method based on the dynamic receptive field comprises the following steps:
the two-dimensional Gaussian first-order partial derivative function
classical receptive field initial response value E (x, y, theta) of each pixel point in each direction parameteri)＝I(x,y)*G(x,y,θi) (4)；
Wherein I (x, y) is the gray value of each pixel point;
the optimal response value E (x, y) of the classical receptive field of each pixel point is max (E (x, y, theta)i)|i＝1,2,…Nθ) (5)。
3. The bionic type contour detection method based on the dynamic receptive field as claimed in claim 2, characterized in that:
the non-classical receptive field response value Inh (x, y) ═ E (x, y) × W (x, y) (6) of each pixel point;
normalized difference function of gaussians
Wherein the content of the first and second substances,
wherein | · | purple1Is L1Norm, h (x) max (0, x);
a final contour response value O (x, y) ═ E (x, y) - β · Inh (x, y) (7) of each pixel point;
wherein β is the non-classical receptor antagonistic strength.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911392146.4A CN111080663B (en) | 2019-12-30 | 2019-12-30 | Bionic contour detection method based on dynamic receptive field |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911392146.4A CN111080663B (en) | 2019-12-30 | 2019-12-30 | Bionic contour detection method based on dynamic receptive field |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111080663A CN111080663A (en) | 2020-04-28 |
| CN111080663B true CN111080663B (en) | 2020-09-22 |
Family
ID=70319385
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911392146.4A Active CN111080663B (en) | 2019-12-30 | 2019-12-30 | Bionic contour detection method based on dynamic receptive field |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111080663B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111968140B (en) * | 2020-06-24 | 2023-07-14 | 广西科技大学 | Contour detection method based on classical receptive field vision-strengthening micro-motion mechanism |
| CN113223009B (en) * | 2021-04-16 | 2023-10-17 | 北京戴纳实验科技有限公司 | Clothes-wearing detection system |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1387315B1 (en) * | 2002-06-07 | 2007-09-12 | Honda Research Institute Europe GmbH | Preparation of a digital image with subsequent edge detection |
| WO2012061669A3 (en) * | 2010-11-05 | 2012-07-12 | Cytognomix,Inc. | Centromere detector and method for determining radiation exposure from chromosome abnormalities |
| CN103903251A (en) * | 2012-12-30 | 2014-07-02 | 南京理工大学 | Night vision image salient contour extracting method based on non-classical receptive field composite modulation |
| CN103927759A (en) * | 2014-04-30 | 2014-07-16 | 武汉大学 | Automatic cloud detection method of aerial images |
| CN104484667A (en) * | 2014-12-30 | 2015-04-01 | 华中科技大学 | Contour extraction method based on brightness characteristic and contour integrity |
| CN106485724A (en) * | 2016-09-20 | 2017-03-08 | 华中科技大学 | A kind of profile testing method that modulates based on combination receptive field and towards feature |
| CN107067408A (en) * | 2017-04-11 | 2017-08-18 | 广西科技大学 | Simulate the image outline detection method of human eye fine motion |
| CN107767387A (en) * | 2017-11-09 | 2018-03-06 | 广西科技大学 | Profile testing method based on the global modulation of changeable reception field yardstick |
| CN107766866A (en) * | 2017-11-07 | 2018-03-06 | 广西科技大学 | Set direction profile testing method based on receptive field subregion |
| CN107909593A (en) * | 2017-11-07 | 2018-04-13 | 广西科技大学 | Non- set direction profile testing method based on receptive field subregion |
| CN108090492A (en) * | 2017-11-09 | 2018-05-29 | 广西科技大学 | The profile testing method inhibited based on scale clue |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014171375A1 (en) * | 2013-04-18 | 2014-10-23 | 旭硝子株式会社 | Glass plate manufacturing method, glass plate manufacturing device and glass plate |
| CN104881851A (en) * | 2015-05-26 | 2015-09-02 | 河海大学 | Adaptive image target enhancement method based on difference of Gaussian model |
| CN106228547B (en) * | 2016-07-15 | 2018-12-28 | 华中科技大学 | A kind of profile and border detection algorithm of view-based access control model color theory and homogeneity inhibition |
| CN110378920B (en) * | 2019-07-17 | 2022-10-14 | 福建师范大学 | Image contour extraction method based on visual model |
-
2019
- 2019-12-30 CN CN201911392146.4A patent/CN111080663B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1387315B1 (en) * | 2002-06-07 | 2007-09-12 | Honda Research Institute Europe GmbH | Preparation of a digital image with subsequent edge detection |
| WO2012061669A3 (en) * | 2010-11-05 | 2012-07-12 | Cytognomix,Inc. | Centromere detector and method for determining radiation exposure from chromosome abnormalities |
| CN103903251A (en) * | 2012-12-30 | 2014-07-02 | 南京理工大学 | Night vision image salient contour extracting method based on non-classical receptive field composite modulation |
| CN103927759A (en) * | 2014-04-30 | 2014-07-16 | 武汉大学 | Automatic cloud detection method of aerial images |
| CN104484667A (en) * | 2014-12-30 | 2015-04-01 | 华中科技大学 | Contour extraction method based on brightness characteristic and contour integrity |
| CN106485724A (en) * | 2016-09-20 | 2017-03-08 | 华中科技大学 | A kind of profile testing method that modulates based on combination receptive field and towards feature |
| CN107067408A (en) * | 2017-04-11 | 2017-08-18 | 广西科技大学 | Simulate the image outline detection method of human eye fine motion |
| CN107766866A (en) * | 2017-11-07 | 2018-03-06 | 广西科技大学 | Set direction profile testing method based on receptive field subregion |
| CN107909593A (en) * | 2017-11-07 | 2018-04-13 | 广西科技大学 | Non- set direction profile testing method based on receptive field subregion |
| CN107767387A (en) * | 2017-11-09 | 2018-03-06 | 广西科技大学 | Profile testing method based on the global modulation of changeable reception field yardstick |
| CN108090492A (en) * | 2017-11-09 | 2018-05-29 | 广西科技大学 | The profile testing method inhibited based on scale clue |
Non-Patent Citations (2)
| Title |
|---|
| Advanced non small cell lung cancer at the National Cancer Institute in Egypt: A descreptive analysis;Hala Aziz Shokralla等;《2 nd International Conference on Respiratory and Pulmonary Medicine》;20161018;第6卷(第5期);全文 * |
| 基于主视通路层级响应模型的轮廓检测方法;周涛等;《航天医学与医学工程》;20180630;第31卷(第3期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111080663A (en) | 2020-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11417148B2 (en) | Human face image classification method and apparatus, and server | |
| CN109117826B (en) | Multi-feature fusion vehicle identification method | |
| CN1977286B (en) | Object recognition method and apparatus therefor | |
| CN106682569A (en) | Fast traffic signboard recognition method based on convolution neural network | |
| CN107767387B (en) | Contour detection method based on variable receptive field scale global modulation | |
| CN105654066A (en) | Vehicle identification method and device | |
| CN111080663B (en) | Bionic contour detection method based on dynamic receptive field | |
| CN108021908B (en) | Face age group identification method and device, computer device and readable storage medium | |
| CN106203461B (en) | Image processing method and device | |
| CN106650617A (en) | Pedestrian abnormity identification method based on probabilistic latent semantic analysis | |
| CN107862267A (en) | Face recognition features' extraction algorithm based on full symmetric local weber description | |
| CN111091075A (en) | Face recognition method and device, electronic equipment and storage medium | |
| Premaratne et al. | Image matching using moment invariants | |
| CN110472693B (en) | Image processing and classifying method and system | |
| Xiang et al. | Moving object detection and shadow removing under changing illumination condition | |
| CN111179294B (en) | Bionic type contour detection method based on X, Y parallel visual channel response | |
| Mohamed et al. | Avatar face recognition using wavelet transform and hierarchical multi-scale LBP | |
| CN111144425B (en) | Method and device for detecting shot screen picture, electronic equipment and storage medium | |
| CN111047618B (en) | Multi-scale-based non-reference screen content image quality evaluation method | |
| CN107194364B (en) | Huffman-L BP multi-pose face recognition method based on divide and conquer strategy | |
| CN114331946A (en) | Image data processing method, device and medium | |
| CN110188646B (en) | Human ear identification method based on fusion of gradient direction histogram and local binary pattern | |
| CN114155572A (en) | Facial expression recognition method and system | |
| Rahmi et al. | Offline signature recognition using back propagation neural network | |
| CN111179293B (en) | Bionic contour detection method based on color and gray level feature fusion |
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 | ||
| EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20200428 Assignee: HUALI FAMILY PRODUCTS CO.,LTD. Assignor: GUANGXI University OF SCIENCE AND TECHNOLOGY Contract record no.: X2023980054119 Denomination of invention: A biomimetic contour detection method based on dynamic receptive field Granted publication date: 20200922 License type: Common License Record date: 20231226 |
|
| EE01 | Entry into force of recordation of patent licensing contract |