CN114266797A - Hammer identification method and device based on probabilistic program inference - Google Patents
Hammer identification method and device based on probabilistic program inference Download PDFInfo
- Publication number
- CN114266797A CN114266797A CN202111493200.1A CN202111493200A CN114266797A CN 114266797 A CN114266797 A CN 114266797A CN 202111493200 A CN202111493200 A CN 202111493200A CN 114266797 A CN114266797 A CN 114266797A
- Authority
- CN
- China
- Prior art keywords
- hammer
- program
- contour
- probability
- shape
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 210000000078 claw Anatomy 0.000 claims abstract description 27
- 238000003708 edge detection Methods 0.000 claims abstract description 21
- 238000000605 extraction Methods 0.000 claims description 8
- 238000001514 detection method Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 4
- 230000000877 morphologic effect Effects 0.000 claims description 4
- 230000002093 peripheral effect Effects 0.000 claims description 4
- 238000010845 search algorithm Methods 0.000 claims description 4
- 238000000691 measurement method Methods 0.000 claims description 3
- 241000245165 Rhododendron ponticum Species 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 9
- 230000008569 process Effects 0.000 description 7
- 230000010429 evolutionary process Effects 0.000 description 5
- 241000251131 Sphyrna Species 0.000 description 4
- 238000013461 design Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 241000208422 Rhododendron Species 0.000 description 3
- 238000009776 industrial production Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the invention provides a hammer identification method and device based on probability program inference. The method comprises the following steps: s1: aiming at a hammer image, extracting contour data of the hammer by adopting an edge detection method, and converting the contour data into a data point set; s2: designing corresponding probability programs aiming at the claw hammer, the stonemason hammer and the vice hammer respectively; s3: and deducing the same hammer image by using a claw hammer probability program, a stonemason hammer probability program and a vice hammer probability program respectively to obtain a claw hammer candidate shape, a stonemason hammer candidate shape and a vice hammer candidate shape, and selecting the shape with the maximum similarity from the candidate shapes as a final shape. The hammer identification method based on probabilistic program inference obtains the hammer outline by using an edge detection method and utilizes probabilistic program inference to identify, is not easily interfered by noise, and can improve the accuracy and the reliability of identification.
Description
Technical Field
The embodiment of the invention relates to the technical field of image recognition, in particular to a hammer recognition method and device based on probability program inference.
Background
At present, the hammer image recognition is mainly applied to the fields related to industrial production and public safety, such as a luggage security inspection system, a hammer production and classification system, and the like, for example: in order to ensure the safety of other public places with a large number of people, such as airports, railway stations, subway stations and the like, the detection of dangerous articles in luggage (including the detection of sharp hammers) becomes one of necessary safety measures; the automatic equipment with good performance and high production efficiency is particularly important in industrial production, and different types of hammers can be accurately conveyed to corresponding production lines through image recognition. The existing method for detecting the hammer through image recognition does not achieve the best effect.
Disclosure of Invention
Based on the above, the embodiment of the invention provides a hammer identification method and device based on probabilistic program inference, so as to improve the accuracy and reliability of hammer identification. The technical scheme is as follows:
in a first aspect, an embodiment of the present application provides a hammer implement identification method based on probabilistic program inference, including the following steps:
s1: aiming at a hammer image, extracting contour data of the hammer by adopting an edge detection method, and converting the contour data into a data point set;
s2: corresponding probability program is designed respectively for claw hammer, stonemason hammer and vice hammer
S3: and deducing by using a claw hammer probability program, a stonemason hammer probability program and a vice hammer probability program respectively aiming at the data point set to obtain claw hammer candidate shapes, stonemason hammer candidate shapes and vice hammer candidate shapes, and selecting the shape with the maximum similarity from the candidate shapes as a final shape.
Optionally, the step of extracting the hammer contour by the edge detection method in S1 is as follows:
s1-1: compressing the read-in image and converting the image into a single-channel gray scale image;
s1-2: carrying out edge detection by using an improved Canny detection algorithm to identify a real edge;
s1-3: removing noise: after the edge detection is finished, eliminating small noise by using morphological 'on' operation, and then reconstructing a part losing the edge by using 'off' operation to fill internal control of a target object to obtain the outline of a main characteristic object;
s1-4: contour extraction: each contour in the image corresponds to a point set, whether the contour is stored or not is determined according to whether the starting point of the contour is closed or not, and then a contour tree is constructed. The contour construction is realized by using functions findContours provided by OpenCV, a RETR _ EXTERNAL retrieval mode is selected for detecting only the outer contour, and a CHAIN _ APPROX _ NONE approximation method is selected for storing all the peripheral contour points.
Optionally, the construction method of the probability program in S2 is as follows:
the number of control points of the hammer head is variable, the number of the control points of the hammer handle is fixed, a cubic B-Spline (B-Spline) method is used for describing Spline curve approximate fitting, and the probability program is described by a closed Spline curve of a plurality of control points.
Optionally, the step of probabilistic program inference in S3 is as follows:
s3-1: running a probability program according to the current parameter value and obtaining a corresponding geometric shape;
s3-2: calculating the similarity between the geometric shape and the image data point set by using an average measurement method;
s3-3: and updating the parameter value of the probability program by utilizing a rhododendron search algorithm, and repeating the steps until the similarity between the geometric shape and the image data point set is basically unchanged.
In a second aspect, the present application provides a hammer implement identification device based on probabilistic program inference, the device including:
the extraction module is used for extracting the contour data of the hammer by adopting an edge detection method aiming at one hammer image and converting the contour data into a data point set;
the program design module is used for designing corresponding probability programs for the claw hammer, the stonemason hammer and the vice hammer respectively;
and the identification module is used for deducing the data point set by using a claw hammer probability program, a stonemason hammer probability program and a vice hammer probability program respectively to obtain a claw hammer candidate shape, a stonemason hammer candidate shape and a vice hammer candidate shape, and selecting a shape with the maximum similarity from the candidate shapes as a finally determined shape.
In order to improve the accuracy and reliability of hammer identification, the invention designs a hammer identification method based on probabilistic program inference. The key to this approach is the use of a hammerhead similarity metric (called the average metric) based on the error from geometry to dataset, which is robust to evaluate the similarity between geometry and an imperfect (noisy, outlier or incomplete) dataset.
The principle and the beneficial effects of the technical scheme are as follows:
the invention extracts the contour data of the hammer image by using an edge detection method, designs corresponding probability programs aiming at a claw hammer, a stonemason hammer and a vice hammer respectively, then deduces the same hammer image by using the claw hammer probability program, the stonemason hammer probability program and the vice hammer probability program respectively to obtain a claw hammer candidate shape, a stonemason hammer candidate shape and a vice hammer candidate shape, and then selects a shape with the maximum similarity from the candidate shapes as a final shape. One of the key steps is probabilistic procedure inference, which is based on the principle that the optimized geometric shape is continuously updated by using a rhododendron search algorithm, so that the contour of the hammer can be completely detected. Compared with other prior art, the technology has the advantages of high data utilization efficiency, good interpretability and strong robustness.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
FIG. 1 is a flow chart illustrating a method for hammer recognition based on probabilistic procedural inference, in accordance with an embodiment of the present invention;
FIGS. 2a-2d are schematic diagrams illustrating the contour extraction of a hammer implement by the edge detection method according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a probabilistic process inference flow shown in an embodiment of the present invention;
FIG. 4 is a schematic flow chart illustrating a method for identifying a hammer implement according to an embodiment of the present disclosure;
FIGS. 5a-5f are graphs of test results provided by embodiments of the present invention;
6a-6c are graphs of similarity evolutionary processes inferred by a hammer probability program provided by embodiments of the present invention;
fig. 7 is a schematic structural diagram of a hammer identification device based on probabilistic procedure inference, which is shown in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In systems such as baggage item security inspection, industrial production classification, etc., hammer identification is often required. The good hammer recognition scheme is beneficial to solving the problems of low detection efficiency, missing detection and the like. The hammer recognition method based on probability program inference provided by the embodiment of the application obtains the hammer outline by using an edge detection method and utilizes probability program inference to recognize the hammer outline, the hammer outline is not easily interfered by noise, and a relatively complete outline can be output, so that the recognition accuracy and reliability are improved, even if a data point set contains a large amount of noise, the method still can keep good recognition capability, and the program has good robustness.
Referring to FIG. 1, the method includes the following steps S1-S3:
s1: aiming at a hammer image, extracting contour data of the hammer by adopting an edge detection method, and converting the contour data into a data point set;
in this application, in step S1, the step of extracting the hammer profile by the edge detection method specifically includes the following steps:
s1-1: compressing the read-in image and converting the image into a single-channel gray scale image;
s1-2: carrying out edge detection by using an improved Canny detection algorithm to identify a real edge;
s1-3: removing noise: after the edge detection is finished, eliminating small noise by using morphological 'on' operation, and then reconstructing a part losing the edge by using 'off' operation to fill internal control of a target object to obtain the outline of a main characteristic object;
s1-4: contour extraction: each contour in the image corresponds to a point set, whether the contour is stored or not is determined according to whether the starting point of the contour is closed or not, and then a contour tree is constructed. The method realizes the construction of the contour by using a function findContours provided by OpenCV, and since only the outer contour of a hammer is needed, a RETR _ EXTERNAL retrieval mode is selected for detecting only the outer contour, and a CHAIN _ APPROX _ NONE approximation method is selected for storing all peripheral contour points.
S2: corresponding probability program is designed respectively for claw hammer, stonemason hammer and vice hammer
The probability programming of the hammer is firstly based on the geometrical characteristics of the hammer, and according to experience and analysis of image contour characteristics, the following characteristics are considered: the hammer can be generally divided into a hammer head and a hammer handle which have larger outline differences; the hammer handle and the hammer head have certain curve characteristics, but the peripheral rectangle can be described by referring to a voxelization method basically and is used for describing the outline of a foundation; the main difference of different hammers is in the hammer head part, and the difference of the hammer handle is smaller. In accordance with the above features, the hammer program can be described by a closed spline curve of a plurality of control points, while a probabilistic program is required to identify and describe how to determine the spline's control points. In a specific implementation, a cubic B-Spline (B-Spline) method is used to describe the Spline curve for the approximate fit. In addition, because the hammer head is the center of the identification hammer, the number of control points for arranging the hammer head is variable, and the number of control points of the hammer handle is fixed in the realization of the text.
The probability program for a hammer has 4 rules: hammer, hammer head, hammer handle and render are shown in tables 1, 2, 3 and 4 below. The program first calls the hammer rule (table 1) to obtain global basic parameters including anchor point I, width W, height H, and the number K of control points of the cubic spline curve. Then, a hammerhead rule (table 2) is called according to the acquired basic parameters, and corresponding K control points are generated in the hammerhead rule, and the function of the control points is to describe a curve of the hammerhead. The shank rule (Table 3) is then invoked based on the anchor point and the underlying rectangle parameters, W and H, which will generate a fixed number (set here to 5) of control points based on the parameters for describing the shank with a simpler shape and structure. Finally, the set of curve points generated after the control points are calculated by interpolation is linearly transformed using the rendering rule (table 4), including rotation and radial transformation (translation has been achieved by the anchor point variables in the hammer rule). And finally, representing the hammer model according to the generated point cloud. Through the analysis of the parameter rule, the total number of parameters which can be represented by the probability program is 8+ K, and the specific variable number of the hammer model is obtained through priori knowledge or training.
TABLE 1 hammer rules
TABLE 2 hammer rules
TABLE 3 hammer shank rules
TABLE 4 rendering rules
S3: and (4) deducing the contour data point set extracted from the hammer image in the S1 by using a claw hammer probability program, a stonemason hammer probability program and a vice hammer probability program respectively to obtain a claw hammer candidate shape, a stonemason hammer candidate shape and a vice hammer candidate shape, and selecting a shape with the maximum similarity from the candidate shapes as a final shape.
Exemplarily, as shown in fig. 4, a hammer image to be subjected to hammer recognition is obtained, n probability programs are respectively used for processing to obtain n (n ≧ 3) candidate shapes, for example, the hammer probability program 1 is used for processing to obtain the candidate shape 1, the hammer probability program 2 is used for processing to obtain the candidate shape 2 …, the obtained n candidate shapes are compared and analyzed, and the candidate shape with the largest similarity is selected as the optimal property, that is, the determined final shape.
In this embodiment of the present application, the probabilistic procedure inference in step S3 includes the following steps:
s3-1: running a probability program according to the current parameter value and obtaining a corresponding geometric shape;
s3-2: calculating the similarity between the geometric shape and the image data point set by using a similarity calculation method (the invention adopts an average measurement method);
s3-3: and updating the parameter values of the probability program by using an optimization algorithm (the invention adopts a rhododendron search algorithm), and repeating the steps until the similarity between the geometric shape and the image data point set is basically unchanged. .
As shown in fig. 3, the process of probabilistic procedure inference roughly includes: inputting a data point set, inputting a probability program and initializing related parameters, obtaining parameters (which can be designated or randomly generated during initialization) according to an optimization algorithm and operating the probability program to generate a corresponding geometric model, calculating the similarity between the current model and the data point set through an estimator, updating parameter values according to a proposed density function of the optimization algorithm, iterating the steps until a target function is converged, and otherwise, adopting the updated parameter values to continue operating the probability program.
Wherein, the extraction process of the hammer image contour point set comprises the following steps: edge detection (see fig. 2 a); morphological operations (see fig. 2 b); contour extraction (see fig. 2 c); the set of contour points obtained (see fig. 2 d).
FIGS. 5a-5f are schematic diagrams of a hammer tool image and a fitting process, wherein FIGS. 5a-5 c are schematic diagrams of a claw hammer, a masonry hammer, and a bench hammer tool image, respectively; fig. 5 d-5 f are the fitting process on the pictures of fig. 5a-5 c using the claw hammer, stonemason and jawshi hammer probability programs, respectively, where the blue part is the data point set extracted from the hammer image and the green part is the fitted geometry.
FIGS. 6a-6c are graphs of the similarity evolutionary process inferred by the probabilistic program of hammers. In the figure, M1, M2 and M3 represent claw hammer, stonemason hammer and vice hammer probability programs respectively. (a) Is a similarity evolutionary process diagram deduced by a probability program on a horn hammer image (see figure 6a) by using M1, M2 and M3; (b) is a similarity evolutionary process diagram deduced by a probability program on a stonemason image (see fig. 6b) by using M1, M2 and M3; (c) is a similarity evolutionary process diagram deduced by a probabilistic program on a jawseed hammer image (see fig. 6c) by using M1, M2 and M3. As can be seen from the similarity evolution process graphs, different types of hammers can be correctly identified.
FIG. 7 is a schematic diagram of a probabilistic inference based hammer recognition apparatus according to an embodiment of the present invention; referring to fig. 7, the hammer recognition device includes:
an extracting module 701, configured to extract, for a hammer image, profile data of the hammer by using an edge detection method, and convert the profile data into a data point set;
a program design module 702, configured to design corresponding probability programs for the claw hammer, the masonry hammer, and the bench hammer, respectively;
and the identification module 703 is configured to perform inference on the data point set by using a claw hammer probability program, a stonemason hammer probability program, and a vice hammer probability program, respectively, to obtain a claw hammer candidate shape, a stonemason hammer candidate shape, and a vice hammer candidate shape, and then select a shape with the largest similarity from the candidate shapes as a finally determined shape.
In summary, the core of the recognition method implemented by the technical solution is to design a robust similarity estimator, a suitable optimization algorithm, and how to properly preprocess image data and write a proper probability program.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (5)
1. A hammer implement identification method based on probabilistic procedural inference, the method comprising the steps of:
s1: aiming at a hammer image, extracting contour data of the hammer by adopting an edge detection method, and converting the contour data into a data point set;
s2: designing corresponding probability programs aiming at the claw hammer, the stonemason hammer and the vice hammer respectively;
s3: and deducing by using a claw hammer probability program, a stonemason hammer probability program and a vice hammer probability program respectively aiming at the data point set to obtain claw hammer candidate shapes, stonemason hammer candidate shapes and vice hammer candidate shapes, and selecting the shape with the maximum similarity from the candidate shapes as the finally determined shape.
2. The method of claim 1, wherein the step of extracting the hammer profile by the edge detection method in S1 is as follows:
s1-1: compressing the read-in image and converting the image into a single-channel gray scale image;
s1-2: carrying out edge detection by using an improved Canny detection algorithm to identify a real edge;
s1-3: removing noise: after the edge detection is finished, eliminating small noise by using morphological 'on' operation, and then reconstructing a part losing the edge by using 'off' operation to fill internal control of a target object to obtain the outline of a main characteristic object;
s1-4: contour extraction: each contour in the image corresponds to a point set, whether the contour is stored or not is determined according to whether the starting point of the contour is closed or not, and then a contour tree is constructed; the contour construction is realized by using functions findContours provided by OpenCV, a RETR _ EXTERNAL retrieval mode is selected for detecting only the outer contour, and a CHAIN _ APPROX _ NONE approximation method is selected for storing all the peripheral contour points.
3. The method according to claim 1, wherein the probabilistic program in S2 is constructed as follows:
the number of control points of the hammer head is variable, the number of the control points of the hammer handle is fixed, a cubic B-Spline (B-Spline) method is used for describing Spline curve approximate fitting, and the probability program is described by a closed Spline curve of a plurality of control points.
4. The method according to claim 1, wherein the step of probabilistic procedure inference in S3 is as follows:
s3-1: running a probability program according to the current parameter value and obtaining a corresponding geometric shape;
s3-2: calculating the similarity between the geometric shape and the image data point set by using an average measurement method;
s3-3: and updating the parameter values of the probability program by using a rhododendron search algorithm, and repeating the steps S3-1 and S3-2 until the similarity between the geometric shape and the image data point set is basically unchanged.
5. A hammer implement identification apparatus based on probabilistic procedural inference, the apparatus comprising:
the extraction module is used for extracting the contour data of the hammer by adopting an edge detection method aiming at one hammer image and converting the contour data into a data point set;
the program design module is used for designing corresponding probability programs for the claw hammer, the stonemason hammer and the vice hammer respectively;
and the identification module is used for deducing the data point set by using a claw hammer probability program, a stonemason hammer probability program and a vice hammer probability program respectively to obtain a claw hammer candidate shape, a stonemason hammer candidate shape and a vice hammer candidate shape, and selecting a shape with the maximum similarity from the candidate shapes as a finally determined shape.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111493200.1A CN114266797A (en) | 2021-12-08 | 2021-12-08 | Hammer identification method and device based on probabilistic program inference |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111493200.1A CN114266797A (en) | 2021-12-08 | 2021-12-08 | Hammer identification method and device based on probabilistic program inference |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114266797A true CN114266797A (en) | 2022-04-01 |
Family
ID=80826607
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111493200.1A Pending CN114266797A (en) | 2021-12-08 | 2021-12-08 | Hammer identification method and device based on probabilistic program inference |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114266797A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090060299A1 (en) * | 2007-08-31 | 2009-03-05 | Hibbard Lyndon S | Method and Apparatus for Efficient Three-Dimensional Contouring of Medical Images |
| CN102799884A (en) * | 2012-07-13 | 2012-11-28 | 北京工商大学 | Image target outline extracting method based on Gestalt rule |
| WO2013084731A1 (en) * | 2011-12-08 | 2013-06-13 | 日本電気株式会社 | Image identifying system |
| CN105787524A (en) * | 2014-12-26 | 2016-07-20 | 中国科学院沈阳自动化研究所 | License plate identification method based on OpenCV and license plate identification system based on OpenCV |
| WO2020064323A1 (en) * | 2018-09-26 | 2020-04-02 | Safran | Method and system for the non-destructive testing of an aerospace part by contour readjustment |
-
2021
- 2021-12-08 CN CN202111493200.1A patent/CN114266797A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090060299A1 (en) * | 2007-08-31 | 2009-03-05 | Hibbard Lyndon S | Method and Apparatus for Efficient Three-Dimensional Contouring of Medical Images |
| WO2013084731A1 (en) * | 2011-12-08 | 2013-06-13 | 日本電気株式会社 | Image identifying system |
| CN102799884A (en) * | 2012-07-13 | 2012-11-28 | 北京工商大学 | Image target outline extracting method based on Gestalt rule |
| CN105787524A (en) * | 2014-12-26 | 2016-07-20 | 中国科学院沈阳自动化研究所 | License plate identification method based on OpenCV and license plate identification system based on OpenCV |
| WO2020064323A1 (en) * | 2018-09-26 | 2020-04-02 | Safran | Method and system for the non-destructive testing of an aerospace part by contour readjustment |
Non-Patent Citations (2)
| Title |
|---|
| 张宗良: "程序式几何模型拟合的理论及其在模式识别中的应用研究", 中国博士学位论文全文数据库 (信息科技辑), 15 June 2021 (2021-06-15), pages 1 - 95 * |
| 张宗良: "程序式几何模型拟合的理论及其在模式识别中的应用研究", 中国博士学位论文全文数据库 信息科技辑, 15 June 2021 (2021-06-15), pages 1 - 95 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110569901B (en) | Channel selection-based countermeasure elimination weak supervision target detection method | |
| Aldoma et al. | Multimodal cue integration through hypotheses verification for rgb-d object recognition and 6dof pose estimation | |
| Myers et al. | Affordance detection of tool parts from geometric features | |
| Martins et al. | Crater detection by a boosting approach | |
| US9483707B2 (en) | Method and device for recognizing a known object in a field of view of a three-dimensional machine vision system | |
| CN110659582A (en) | Image conversion model training method, heterogeneous face recognition method, device and equipment | |
| US20160196467A1 (en) | Three-Dimensional Face Recognition Device Based on Three Dimensional Point Cloud and Three-Dimensional Face Recognition Method Based on Three-Dimensional Point Cloud | |
| US10311635B2 (en) | Method and apparatus for detecting repetitive structures in 3D mesh models | |
| Arbeiter et al. | Evaluation of 3D feature descriptors for classification of surface geometries in point clouds | |
| JP5984327B2 (en) | Information processing method and apparatus, and program | |
| CN111091080A (en) | Face recognition method and system | |
| CN114170184A (en) | Product image anomaly detection method and device based on embedded feature vector | |
| JP2019046094A (en) | Information processing device, control method and program for information processing device | |
| LU503034B1 (en) | Palmprint Recognition Method Based on Fusion Depth Network | |
| CN111582358A (en) | Training method and device for house type recognition model and house type weight judging method and device | |
| CN110826571B (en) | Image traversal algorithm for rapid image identification and feature matching | |
| CN114219753A (en) | Power equipment surface defect detection method based on deep learning and terminal | |
| CN114266797A (en) | Hammer identification method and device based on probabilistic program inference | |
| Wu et al. | An effective texture spectrum descriptor | |
| US11244206B2 (en) | Image normalization for facial analysis | |
| CN111194004B (en) | Base station fingerprint positioning method, device and system and computer readable storage medium | |
| JP3859347B2 (en) | Object recognition apparatus and method | |
| Ragab et al. | Sign language recognition using Hilbert curve features | |
| Laiche et al. | Geometric based histograms for shape representation and retrieval | |
| Howarth et al. | Feature-based object recognition |
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 |