CN111191659B - Multi-shape clothes hanger identification method for clothing production system - Google Patents
Multi-shape clothes hanger identification method for clothing production system Download PDFInfo
- Publication number
- CN111191659B CN111191659B CN201911367810.XA CN201911367810A CN111191659B CN 111191659 B CN111191659 B CN 111191659B CN 201911367810 A CN201911367810 A CN 201911367810A CN 111191659 B CN111191659 B CN 111191659B
- Authority
- CN
- China
- Prior art keywords
- image
- hanger
- target
- color
- clothes
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/457—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by analysing connectivity, e.g. edge linking, connected component analysis or slices
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention discloses a method for identifying multi-shape clothes hangers in a clothes production system, which comprises the steps of firstly, acquiring clothes hanger images through a camera, preprocessing the acquired clothes hanger images and extracting edges to acquire target clothes hangers; the edge information of the target clothes hanger is complemented by a closed filter; clustering by adopting a K-means method to divide the target clothes hanger; respectively extracting an edge image of a target clothes hanger and an image clustered by K-means; extracting the outer contour of an edge map of the target hanger by using a boundary acquisition tracking method; the outer contour of the clothes hanger is used as a filling boundary, and the shapes of clothes hangers with various shapes are filled by using a water-diffusion filling method. And storing the extracted hangers of different shapes into a database, and matching according to Euclidean distance to realize autonomous identification of the hangers of different shapes. The invention solves the problem of low efficiency caused by manual identification in the transmission process of clothes hangers of different shapes under the complex background of the existing clothing production system.
Description
Technical Field
The invention belongs to the technical field of clothes hanger detection and machine vision, and particularly relates to a multi-shape clothes hanger identification method of a clothes production system.
Background
Along with the strict requirements of the clothing processing industry on production efficiency, the identification of the clothes hanger in a complex background is one of important processes in the production of clothing production lines. In particular, during the movement process of clothes hangers with different shapes, the clothes hangers are detected and classified by manpower, so that the processing efficiency is seriously affected. Therefore, the automatic identification research on clothes hangers of different shapes is realized, the production efficiency of a clothing processing line can be improved, and the development of automation is indirectly promoted.
How to improve the identification and classification of different types of hangers in motion in complex production workshops is one of the effective ways to increase garment production efficiency. In recent years, image processing technology and intelligent system terminal equipment are applied in many fields, and therefore, it has become a trend to perform hanger identification by using an identification method with strong identification capability and high accuracy. The target clothes hanger identification method based on the image processing technology can be conveniently applied to a microprocessor terminal, and the power consumption and the cost are relatively low. Therefore, the visual method is of a certain significance in identifying the clothing hangers.
Disclosure of Invention
The invention aims to provide a multi-shape clothes hanger identification method for a clothes production system, which solves the problem of low efficiency caused by manual identification in the transmission process of clothes hangers of different shapes under the complex background of the existing clothes production system.
The technical scheme adopted by the invention is that the method for identifying the multi-shape clothes hangers of the clothes production system is implemented according to the following steps:
step 1, acquiring an image of a clothes hanger through a camera, and preprocessing to enhance the effective characteristics of the image;
step 2, carrying out color channel ratio image transformation on the preprocessed clothes hanger image, and extracting the edge of the target clothes hanger by selecting two colors to obtain an edge image of the target clothes hanger;
step 3, connecting the missing edge information in the extracted hanger edge image to form a complete hanger edge image by closing the filter;
step 4, adopting a K-means clustering method to the hanger image obtained in the step 2, dividing the image into a region close to the color of the hanger and a region with large deviation from the color of the hanger, respectively selecting a plurality of initial clustering centers in the two regions, calculating the distance from each pixel point of each region in the image to the initial clustering center of the region, selecting the class of the closest clustering center of the hanger to be identified according to the change of the distance range parameter in the clustering method, taking the class as the region close to the color of the hanger, namely the target hanger region, dividing the target hanger, and forming a target hanger region image;
step 5, fusing the images obtained in the step 3 and the step 4 to obtain the edges and the areas of the target clothes hangers in the images;
step 6, extracting the outline of the edge image of the target clothes hanger by using a method of collecting boundary tracking;
step 7, using the image obtained by the K-means clustering method in the step 4 as a central area of the target clothes hanger, taking the outer contour of the clothes hanger obtained in the step 6 as a filling boundary, and filling the shapes of the clothes hangers with various shapes by using a water-diffusion filling method;
and 8, storing the clothes hangers of different shapes extracted by the method into a database to serve as templates, and matching the templates with images acquired in real time on a production site and extracted according to the method according to Euclidean distance to realize autonomous identification of the clothes hangers of different shapes.
The present invention is also characterized in that,
and step 1, judging whether the image is a color cast image according to the input image when the image is preprocessed, and if the image is the color cast image, correcting the color cast image to obtain a color cast corrected image. The implementation process for correcting the color cast comprises the following steps:
step 1.1, set image f (x) = [ f r (x),f g (x),f b (x)] T Wherein x represents pixel coordinates, and the dynamic range of corresponding x is [0, L]L is the largest pixel of the image, f r (x)、f g (x)、f b (x) For three color channels, for each channel f c (x) C=r, g, b, the image histogram is represented as a 2×n matrix:
wherein N is the column number of the matrix, H C For the image histogram matrix, h cN A vector of N gray levels, p cN For vectors composed of probabilities corresponding to N gray levels, solving a histogram matrix of an original image through a formula (1.1), and solving a maximum pixel L of the histogram matrix;
in step 1.2, in order to make the distribution of the image histogram approximate to uniform distribution, the distance between two adjacent gray levels in the histogram in step 1.1 is calculated as follows:
wherein s is cn H is the distance between two adjacent gray levels cn H is the current gray level c,n-1 And for the gray level of the previous stage, the effective characteristic information of the image is finally enhanced through algorithm calculation, and optimization is provided for further edge extraction of the image.
The color channel ratio image transformation in step 2 is specifically as follows:
selecting a color channel ratio image according to prior color information of a target, wherein the pixel value of a corresponding region of the target in the ratio image is larger, the comparison is outstanding, the pixel value of other non-target regions is smaller, and the region is restrained, so that two color channels are respectively selected as channels with obvious target colors, marked as foreground channels, the rest color channels are marked as background channels, the selection of an image region is carried out, and when the foreground channels are higher than the background channels, the regions which are expressed as targets and are close to the target colors are displayed; when the foreground channel is approximately equal to the background channel, the foreground channel is represented as a region without obvious color; when the foreground channel is smaller than the background channel, it appears as a region that is significantly different from the target color.
The color clustering process of the clothes rack image in the step 4 is specifically as follows:
step 4.1, firstly, inputting a data set M containing n elements, a class center number k and a threshold value xi; let initial value i=1, select K initial cluster centers C from the data set M with K-means clusters j = (I), j=1, 2, …, k, this initial cluster center being a point within the target hanger area;
step 4.2, calculating the distance D (x) from each sample point in M to the cluster center i ,Z j (I) I=1, 2, …, n, j=1, 2, …, k, if satisfied
Then x i The method belongs to the m-th class, and the maximum and minimum distance values from the sample points in the hanger area to the initial clustering center of the hanger can be obtained through calculation;
step 4.3, calculating errorSum of squares criterion function
Obtaining the distance between each pixel point and the initial clustering center of the area, and further preparing for determining the class of the clustering center for the target hanger;
step 4.4, if|J c (I)-J c If (I-1) | < ζ, ending the algorithm, otherwise, calculating k new clustering centers, repeating the steps 4.2-4.4, finally outputting a set divided into k categories, and comparing the set with data sets stored in a database by hangers of different shapes to obtain a certain same category, wherein the category is a target hanger area image; wherein C is j For the initial cluster center, D (x i ,Z j (I) Distance from each sample point in the data set to the clustering center, m is the number of category centers, J c As an output value of the clustering criterion function,
the multi-shape clothes hanger identification method has the beneficial effects that the identification is carried out under a complex background by fusing the characteristics of the shape outline and the image sample point clustering, and compared with a single background, the identification result is more effective; the problem of low working efficiency in the process of identifying and classifying the traditional different clothes hangers is solved, and the intellectualization of clothing manufacture is improved.
Drawings
FIG. 1 is a color space of a garment hanger in a vision-based garment hanger identification method of the present invention;
fig. 2 is a schematic view of the flood filling of an optimized target area in a vision-based garment hanger identification method of the present invention.
Detailed Description
The invention will be described in detail below with reference to the drawings and the detailed description.
The invention relates to a method for identifying multi-shape clothes hangers in a clothes production system, which is implemented according to the following steps:
step 1, acquiring an image of a clothes hanger through a camera, and preprocessing to enhance the effective characteristics of the image;
in fig. 1, the color distinction degree of the image of the hanger collected by the camera is not high, and the accurate identification of various hangers cannot be achieved, so that the color space of the hanger color is formed by fusing the extracted color of the hanger with the RGB color standard.
Step 2, carrying out color channel ratio image transformation on the preprocessed clothes hanger image, and extracting the edge of the target clothes hanger by selecting two colors to obtain an edge image of the target clothes hanger;
step 3, connecting the missing edge information in the extracted hanger edge image to form a complete hanger edge image by closing the filter;
step 4, adopting a K-means clustering method to the hanger image obtained in the step 2, dividing the image into a region close to the color of the hanger and a region with large deviation from the color of the hanger, respectively selecting a plurality of initial clustering centers in the two regions, calculating the distance from each pixel point of each region in the image to the initial clustering center of the region, selecting the class of the closest clustering center of the hanger to be identified according to the change of the distance range parameter in the clustering method, taking the class as the region close to the color of the hanger, namely the target hanger region, dividing the target hanger, and forming a target hanger region image;
step 5, fusing the images obtained in the step 3 and the step 4 to obtain the edges and the areas of the target clothes hangers in the images;
step 6, extracting the outline of the edge image of the target clothes hanger by using a method of collecting boundary tracking;
step 7, using the image obtained by the K-means clustering method in the step 4 as a central area of the target clothes hanger, taking the outer contour of the clothes hanger obtained in the step 6 as a filling boundary, and filling the shapes of the clothes hangers with various shapes by using a water-diffusion filling method;
fig. 2 is a view of the fact that the complete image cannot be formed due to the fact that the non-uniform extraction is not performed at the place where the features inside the edge of the hanger are fine, so that the non-recognized portions of the outline area of the hanger are filled by the diffused water filling, and the hanger is formed into a complete shape.
And 8, storing the clothes hangers of different shapes extracted by the method into a database to serve as templates, and matching the templates with images acquired in real time on a production site and extracted according to the method according to Euclidean distance to realize autonomous identification of the clothes hangers of different shapes.
When the image is preprocessed in the step 1, judging whether the image is a color cast image or not according to the input image, and if the image is the color cast image, correcting the color cast image to obtain a color cast corrected image. The implementation process for correcting the color cast comprises the following steps:
step 1.1, set image f (x) = [ f r (x),f g (x),f b (x)] T Wherein x represents pixel coordinates, and the dynamic range of corresponding x is [0, L]L is the largest pixel of the image, f r (x)、f g (x)、f b (x) For three color channels, for each channel f c (x) C=r, g, b, the image histogram is represented as a 2×n matrix:
wherein N is the column number of the matrix, H C For the image histogram matrix, h cN A vector of N gray levels, p cN For vectors composed of probabilities corresponding to N gray levels, solving a histogram matrix of an original image through a formula (1.1), and solving a maximum pixel L of the histogram matrix;
in step 1.2, in order to make the distribution of the image histogram approximate to uniform distribution, the distance between two adjacent gray levels in the histogram in step 1.1 is calculated as follows:
wherein s is cn H is the distance between two adjacent gray levels cn H is the current gray level c,n-1 For the previous gray level, calculated by algorithmAnd finally, enhancing the effective characteristic information of the image and providing optimization for the edge extraction of the further image.
The color channel ratio image transformation in step 2 is specifically as follows:
selecting a color channel ratio image according to prior color information of a target, wherein the pixel value of a corresponding region of the target in the ratio image is larger, the comparison is outstanding, the pixel value of other non-target regions is smaller, and the region is restrained, so that two color channels are respectively selected as channels with obvious target colors, marked as foreground channels, the rest color channels are marked as background channels, the selection of an image region is carried out, and when the foreground channels are higher than the background channels, the regions which are expressed as targets and are close to the target colors are displayed; when the foreground channel is approximately equal to the background channel, the foreground channel is represented as a region without obvious color; when the foreground channel is smaller than the background channel, it appears as a region that is significantly different from the target color.
The color clustering process of the clothes rack image in the step 4 is specifically as follows:
step 4.1, firstly, inputting a data set M containing n elements, a class center number k and a threshold value xi; let initial value i=1, select K initial cluster centers C from the data set M with K-means clusters j = (I), j=1, 2, …, k, this initial cluster center being a point within the target hanger area;
step 4.2, calculating the distance D (x) from each sample point in M to the cluster center i ,Z j (I) I=1, 2, …, n, j=1, 2, …, k, if satisfied
Then x i The method belongs to the m-th class, and the maximum and minimum distance values from the sample points in the hanger area to the initial clustering center of the hanger can be obtained through calculation;
step 4.3, calculating the error square sum criterion function
Obtaining the distance from each pixel point to the initial clustering center of the area, and further determining the class of the clustering center for the target hangerPreparing;
step 4.4, if|J c (I)-J c If (I-1) | < ζ, ending the algorithm, otherwise, calculating k new clustering centers, repeating the steps 4.2-4.4, finally outputting a set divided into k categories, and comparing the set with data sets stored in a database by hangers of different shapes to obtain a certain same category, wherein the category is a target hanger area image; wherein C is j For the initial cluster center, D (x i ,Z j (I) Distance from each sample point in the data set to the clustering center, m is the number of category centers, J c As an output value of the clustering criterion function,
the clothing hanger identification method based on vision is characterized by improving the informatization level on a production line, effectively, intelligently, stably and permanently realizing the classification identification of clothing hangers, and simultaneously being applied to a mobile terminal to meet the portable requirement.
Claims (4)
1. The multi-shape clothes hanger identification method of the clothing production system is characterized by comprising the following steps of:
step 1, acquiring an image of a clothes hanger through a camera, and preprocessing to enhance the effective characteristics of the image;
step 2, carrying out color channel ratio image transformation on the preprocessed clothes hanger image, and extracting the edge of the target clothes hanger by selecting two colors to obtain an edge image of the target clothes hanger;
step 3, connecting the missing edge information in the extracted hanger edge image to form a complete hanger edge image by closing the filter;
step 4, adopting a K-means clustering method to the hanger image obtained in the step 2, dividing the image into a region close to the color of the hanger and a region with large deviation from the color of the hanger, respectively selecting a plurality of initial clustering centers in the two regions, calculating the distance from each pixel point of each region in the image to the initial clustering center of the region, selecting the class of the closest clustering center of the hanger to be identified according to the change of the distance range parameter in the clustering method, taking the class as the region close to the color of the hanger, namely the target hanger region, dividing the target hanger, and forming a target hanger region image;
step 5, fusing the images obtained in the step 3 and the step 4 to obtain the edges and the areas of the target clothes hangers in the images;
step 6, extracting the outline of the edge image of the target clothes hanger by using a method of collecting boundary tracking;
step 7, using the image obtained by the K-means clustering method in the step 4 as a central area of the target clothes hanger, taking the outer contour of the clothes hanger obtained in the step 6 as a filling boundary, and filling the shapes of the clothes hangers with various shapes by using a water-diffusion filling method;
and 8, storing the clothes hangers of different shapes extracted by the method into a database to serve as templates, and matching the templates with images acquired in real time on a production site and extracted according to the method according to Euclidean distance to realize autonomous identification of the clothes hangers of different shapes.
2. The method for recognizing the multi-shape clothes hangers in the clothing production system according to claim 1, wherein when the image is preprocessed in the step 1, whether the image is a color cast image is judged according to the input image, and if the image is the color cast image, the color cast image is required to be corrected, so that a color cast corrected image is obtained; the implementation process for correcting the color cast comprises the following steps:
step 1.1, set image f (x) = [ f r (x),f g (x),f b (x)] T Wherein x represents pixel coordinates, and the dynamic range of corresponding x is [0, L]L is the largest pixel of the image, f r (x)、f g (x)、f b (x) For three color channels, for each channel f c (x) C=r, g, b, the image histogram is represented as a 2×n matrix:
wherein N is the column number of the matrix, H C For the image histogram matrix, h cN A vector of N gray levels, p cN For vectors composed of probabilities corresponding to N gray levels, solving a histogram matrix of an original image through a formula (1.1), and solving a maximum pixel L of the histogram matrix;
in step 1.2, in order to make the distribution of the image histogram approximate to uniform distribution, the distance between two adjacent gray levels in the histogram in step 1.1 is calculated as follows:
wherein s is cn H is the distance between two adjacent gray levels cn H is the current gray level c,n-1 And for the gray level of the previous stage, the effective characteristic information of the image is finally enhanced through algorithm calculation, and optimization is provided for further edge extraction of the image.
3. The method for identifying multi-shape hangers in a clothing production system according to claim 1, wherein the color channel ratio image transformation in step 2 is specifically as follows:
selecting a color channel ratio image according to prior color information of a target, wherein the pixel value of a corresponding region of the target in the ratio image is larger, the comparison is outstanding, the pixel value of other non-target regions is smaller, and the region is restrained, so that two color channels are respectively selected as channels with obvious target colors, marked as foreground channels, the rest color channels are marked as background channels, the selection of an image region is carried out, and when the foreground channels are higher than the background channels, the regions which are expressed as targets and are close to the target colors are displayed; when the foreground channel is approximately equal to the background channel, the foreground channel is represented as a region without obvious color; when the foreground channel is smaller than the background channel, it appears as a region that is significantly different from the target color.
4. The method for identifying multi-shape hangers in a clothing production system according to claim 1, wherein the color clustering process of the hanger image in the step 4 is specifically as follows:
step 4.1, firstly, inputting a data set M containing n elements, a class center number k and a threshold value xi; let initial value i=1, select K initial cluster centers C from the data set M with K-means clusters j = (I), j=1, 2, …, k, this initial cluster center being a point within the target hanger area;
step 4.2, calculating the distance D (x) from each sample point in M to the cluster center i ,Z j (I) I=1, 2, …, n, j=1, 2, …, k, if satisfied
Then x i The method belongs to the m-th class, and the maximum and minimum distance values from the sample points in the hanger area to the initial clustering center of the hanger can be obtained through calculation;
step 4.3, calculating the error square sum criterion function J c :
Obtaining the distance between each pixel point and the initial clustering center of the area, and further preparing for determining the class of the clustering center for the target hanger;
step 4.4, if|J c (I)-J c If (I-1) | < ζ, ending the algorithm, otherwise, calculating k new clustering centers, repeating the steps 4.2-4.4, finally outputting a set divided into k categories, and comparing the set with data sets stored in a database by hangers of different shapes to obtain a certain same category, wherein the category is a target hanger area image; wherein C is j For the initial cluster center, D (x i ,Z j (I) Distance from each sample point in the data set to the clustering center, m is the number of category centers, J c For the output value of the clustering criterion function, J c :
/>
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911367810.XA CN111191659B (en) | 2019-12-26 | 2019-12-26 | Multi-shape clothes hanger identification method for clothing production system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911367810.XA CN111191659B (en) | 2019-12-26 | 2019-12-26 | Multi-shape clothes hanger identification method for clothing production system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111191659A CN111191659A (en) | 2020-05-22 |
| CN111191659B true CN111191659B (en) | 2023-04-28 |
Family
ID=70708017
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911367810.XA Active CN111191659B (en) | 2019-12-26 | 2019-12-26 | Multi-shape clothes hanger identification method for clothing production system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111191659B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814877B (en) * | 2020-07-09 | 2023-11-24 | 北京服装学院 | System and method for live-action exhibition of large-scale group performance movable clothing |
| CN114186299B (en) * | 2021-12-09 | 2023-12-15 | 上海百琪迈科技(集团)有限公司 | Method for generating and rendering three-dimensional clothing seam effect |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8861873B2 (en) * | 2010-06-01 | 2014-10-14 | Hewlett-Packard Development Company, L.P. | Image clustering a personal clothing model |
| US20140310304A1 (en) * | 2013-04-12 | 2014-10-16 | Ebay Inc. | System and method for providing fashion recommendations |
| CN108154087A (en) * | 2017-12-08 | 2018-06-12 | 北京航天计量测试技术研究所 | A kind of matched infrared human body target detection tracking method of feature based |
| CN108764062B (en) * | 2018-05-07 | 2022-02-25 | 西安工程大学 | Visual sense-based clothing piece identification method |
-
2019
- 2019-12-26 CN CN201911367810.XA patent/CN111191659B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111191659A (en) | 2020-05-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | A detection method for apple fruits based on color and shape features | |
| CN109636784B (en) | Image saliency target detection method based on maximum neighborhood and super-pixel segmentation | |
| CN111368683B (en) | Face image feature extraction method and face recognition method based on modular constraint CenterFace | |
| CN108537239B (en) | Method for detecting image saliency target | |
| CN106803257B (en) | Method for segmenting disease spots in crop disease leaf image | |
| CN107392968B (en) | The image significance detection method of Fusion of Color comparison diagram and Color-spatial distribution figure | |
| CN102324032B (en) | Texture feature extraction method for gray level co-occurrence matrix in polar coordinate system | |
| CN107230188B (en) | Method for eliminating video motion shadow | |
| CN104881671B (en) | A kind of high score remote sensing image Local Feature Extraction based on 2D Gabor | |
| CN108710916B (en) | Picture classification method and device | |
| CN109145964B (en) | Method and system for realizing image color clustering | |
| CN111191659B (en) | Multi-shape clothes hanger identification method for clothing production system | |
| CN105069816B (en) | A kind of method and system of inlet and outlet people flow rate statistical | |
| CN109886267A (en) | A kind of soft image conspicuousness detection method based on optimal feature selection | |
| CN102208099A (en) | Illumination-change-resistant crop color image segmentation method | |
| CN109472770B (en) | Method for quickly matching image characteristic points in printed circuit board detection | |
| CN109447111A (en) | A kind of remote sensing supervised classification method based on subclass training sample | |
| CN107657276B (en) | Weak supervision semantic segmentation method based on searching semantic class clusters | |
| CN112163636A (en) | Unknown pattern recognition method of electromagnetic signal radiation source based on twin neural network | |
| CN105354547A (en) | Pedestrian detection method in combination of texture and color features | |
| CN113408573B (en) | Method and device for automatically classifying and classifying tile color numbers based on machine learning | |
| CN111127407B (en) | Fourier transform-based style migration forged image detection device and method | |
| CN109427068A (en) | Deep learning based on super-pixel automatic marking is divided into segmentation method | |
| CN112784869A (en) | Fine-grained image identification method based on attention perception and counterstudy | |
| CN104036232A (en) | Image edge feature analysis-based necktie pattern retrieval method |
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 |