CN109829380A - A kind of detection method, device, system and the storage medium of dog face characteristic point - Google Patents
A kind of detection method, device, system and the storage medium of dog face characteristic point Download PDFInfo
- Publication number
- CN109829380A CN109829380A CN201811628345.6A CN201811628345A CN109829380A CN 109829380 A CN109829380 A CN 109829380A CN 201811628345 A CN201811628345 A CN 201811628345A CN 109829380 A CN109829380 A CN 109829380A
- Authority
- CN
- China
- Prior art keywords
- characteristic point
- face
- dog face
- profile
- point
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention provides mask method, device, system and the computer storage mediums of a kind of dog face characteristic point.The mask method of the dog face characteristic point includes: to carry out characteristic point detection based on the image comprising dog face and trained detection model, obtains the characteristic point of fine positioning;Wherein, the detection model includes first order network and second level network, and the characteristic point for obtaining fine positioning includes: that the first order network of full-face images and detection model based on the dog face carries out characteristic point detection, obtains the characteristic point of coarse positioning;The second level network of topography and detection model based on the dog face positions the characteristic point of coarse positioning, obtains the characteristic point of the fine positioning.According to the method for the present invention, device, system and computer storage medium can effectively improve the accuracy and real-time of the detection of dog face characteristic point.
Description
Technical field
The present invention relates to technical field of image processing, relate more specifically to the processing of dog face image.
Background technique
Characteristic point mark largely influences image recognition, analyzes, searches as a very important step before image alignment
The overall performance of cable system.At present in the process of face recognition, there are many effective characteristic point mask methods, but for
The method of characteristic point mark in animal identification is considerably less, such as to the mark of dog face characteristic point in the identification of dog face.
If be labeled using traditional characteristic point mark and detection method to dog face, since each key point is independent
Detection, the global geological information of dog face are almost completely neglected, this makes it very sensitive to subtle disturbance, to illumination variation, appearance
The robustness such as state variation are bad.In addition, the number for calculating time, complexity and characteristic point is directly proportional, characteristic point to be detected
More, required detector is more, this is difficult to carry out it in the application of closeer characteristic point.
Therefore, lack the method being preferably labeled for dog face characteristic point, traditional characteristic point mark in the prior art
Injecting method is very big by subtle disturbing influence, be easy to cause and fails to report or report by mistake, so that accuracy and recall rate are low, when mark point is more
When, cause operational efficiency low.
Summary of the invention
The present invention is proposed in view of the above problem.The present invention provides a kind of mask methods of dog face characteristic point, dress
It sets, system and computer storage medium can be mentioned effectively by the Multilevel ANN based on full face and local message foundation
The accuracy and real-time of high dog face characteristic point detection.
According to an aspect of the present invention, a kind of detection method of dog face characteristic point is provided, comprising:
Characteristic point detection is carried out based on the image comprising dog face and trained detection model, obtains the feature of fine positioning
Point;
Wherein, the detection model includes first order network and second level network, the characteristic point packet for obtaining fine positioning
It includes:
The first order network of full-face images and detection model based on the dog face carries out characteristic point detection, obtains coarse positioning
Characteristic point;
The second level network of topography and detection model based on the dog face positions the characteristic point of coarse positioning,
Obtain the characteristic point of the fine positioning.
Illustratively, the method also includes: the dog face full-face images are split according to the position of dog face,
Obtain the topography of the dog face.
Illustratively, characteristic point of the second level network of topography and detection model based on the dog face to coarse positioning
It is positioned, the characteristic point for obtaining the fine positioning includes:
The second level network of topography and detection model based on the dog face positions the characteristic point of coarse positioning,
Obtain the characteristic point of topography;
The characteristic point of topography is coordinately transformed and is integrated to obtain the characteristic point of the fine positioning of the dog face.
Illustratively, the characteristic point of topography is coordinately transformed and is integrated to obtain the spy of the fine positioning of the dog face
Sign point, comprising:
Obtain reference position and rotation angle of the topography of dog face relative to full-face images;
The characteristic point of corresponding topography is coordinately transformed according to the reference position and rotation angle, is obtained
The characteristic point of transformed topography;
The characteristic point of topography after each transformation is integrated, the characteristic point of the fine positioning of dog face is obtained.
Illustratively, the training of the detection model includes: based on pre-defined rule to training sample full-face images and training
Dog face in sample topography carries out characteristic point mark;
Based on after mark training sample full-face images and training sample topography detection model is trained, obtain
Trained detection model.
Illustratively, the pre-defined rule includes ear's profile, eye profile, nasal contours, mouth based on the dog face
The mark of at least one of contouring, face mask progress characteristic point.
Illustratively, it includes: the left and right side for marking the basal part of the ear that ear's profile based on the dog face, which carries out the mark of characteristic point,
Boundary's characteristic point, basal part of the ear central feature point, have sharp ears characteristic point and on the basis of basal part of the ear central feature point to have sharp ears characteristic point it is equidistant
The characteristic point of mark.
Illustratively, it includes: mark left eye central feature that the eye profile based on the dog face, which carries out the mark of characteristic point,
The characteristic point and left eye central vertical line and institute that point, left eye central horizontal line intersect with the left and right sides of the left eye profile
State the characteristic point of the intersection of two sides up and down of left eye profile;With
The feature that mark right eye central feature point, right eye central horizontal line intersect with the left and right sides of the right eye profile
The characteristic point that point and right eye central vertical line intersect with the two sides up and down of the right eye profile.
Illustratively, the eye profile based on the dog face carries out the mark of characteristic point further include: with the left eye center
On the basis of the characteristic point that horizontal line and left eye central vertical line intersect with the left eye profile, equidistantly marked along the left eye profile
Characteristic point;With
On the basis of the characteristic point that the right eye central horizontal line and right eye central vertical line intersect the right eye profile,
Characteristic point is equidistantly marked along the right eye profile.
Illustratively, it includes: mark nose central feature point that the nasal contours based on the dog face, which are labeled,.
Illustratively, it includes: mark left side corners of the mouth characteristic point, upper lip that the mouth profile based on the dog face, which is labeled,
Left side profile corner feature point, upper lip central feature point, upper lip right lateral contours corner feature point, right side corners of the mouth characteristic point,
Profile corner feature point, lower lip central feature point, lower lip right lateral contours corner feature point on the left of lower lip.
Illustratively, the face mask based on the dog face, which is labeled, includes:
Edema with the heart involved in characteristic point that mark crown central feature point, left eye central horizontal line and face left side profile intersect, nose
The characteristic point and nose of characteristic point, right eye central horizontal line and the intersection of face right lateral contours that horizontal line intersects with profile on the left of face
The characteristic point that central horizontal line intersects with face right lateral contours.
Illustratively, the face mask based on the dog face is labeled further include: with the left eye central horizontal line with
On the basis of the characteristic point that the characteristic point of profile intersection and the nose central horizontal line intersect with profile on the left of face on the left of face, edge
On the left of face profile equidistantly mark characteristic point;The characteristic point intersected with the right eye center line with face right lateral contours and the nose
Point on the basis of the characteristic point that sharp central horizontal line intersects with face right lateral contours equidistantly marks characteristic point along face right lateral contours.
According to a further aspect of the invention, a kind of detection device of dog face characteristic point is provided, comprising:
Detection module is obtained for carrying out characteristic point mark based on the image comprising dog face and trained detection model
The characteristic point of fine positioning;
Wherein, the detection model includes first order network and second level network, and the first order network is used to be based on institute
The first order network of the full-face images and detection model of stating dog face carries out characteristic point detection, obtains the characteristic point of coarse positioning;
Second level network of the second level network for topography and detection model based on the dog face is determined thick
The characteristic point of position is positioned, and the characteristic point of the fine positioning is obtained.
Illustratively, the training of the detection model include: based on pre-defined rule to training sample image full-face images and
Dog face in training sample topography carries out characteristic point mark;
Based on after mark training sample full-face images and training sample topography detection model is trained, obtain
Trained detection model.
Illustratively, the training of the first order network of the detection model includes:
The full face sample image of dog face before being marked based on the training sample image before mark, and based on mark
The training sample image afterwards marked after the full face sample image of dog face;
First nerves network is trained according to the dog face full face sample image after mark, obtains trained detection mould
The first order network of type.
Illustratively, the detection module is also used to: being carried out according to the position of dog face to the dog face full-face images
Segmentation, obtains the topography of the dog face.
Illustratively, the training of the second level network of the detection model includes:
Dog face fractional sample image before being marked based on the training sample image before mark, and based on mark
The training sample image afterwards marked after dog face fractional sample image;
Nervus opticus network is trained according to the dog face fractional sample image after mark, obtains trained detection mould
The second level network of type.
Illustratively, the second level network is further used for:
Topography based on the dog face positions the characteristic point of coarse positioning, obtains the characteristic point of topography;
The characteristic point of topography is coordinately transformed and is integrated, the characteristic point of the fine positioning of the dog face is obtained.
Illustratively, the second level network is further used for:
Obtain reference position and rotation angle of the topography of dog face relative to full-face images;
The characteristic point of corresponding topography is coordinately transformed according to the reference position and rotation angle, is obtained
The characteristic point of transformed topography;
The characteristic point of topography after each transformation is integrated, the characteristic point of the fine positioning of dog face is obtained.
Illustratively, described device further include: output module includes the dog face characteristic point and/or described for exporting
The dog face image of dog face characteristic point coordinate.
According to a further aspect of the invention, provide a kind of detection system of dog face characteristic point, including memory, processor and
It is stored in the computer program run on the memory and on the processor, which is characterized in that the processor executes
The step of above method is realized when the computer program.
According to a further aspect of the invention, a kind of computer readable storage medium is provided, computer program is stored thereon with,
It is characterized in that, the step of realizing the above method when computer program is computer-executed.
Detection method, device, system and the computer storage medium of dog face characteristic point according to an embodiment of the present invention, pass through
Based on the cascade neural network that full face and local message are established, Query refinement is predicted the location information of dog face characteristic point, is realized
The high accuracy positioning of dog face characteristic point can effectively improve the accuracy and real-time of the detection of dog face characteristic point, can be extensive
Applied to the various occasions handled about dog face image.
Detailed description of the invention
The embodiment of the present invention is described in more detail in conjunction with the accompanying drawings, the above and other purposes of the present invention,
Feature and advantage will be apparent.Attached drawing is used to provide to further understand the embodiment of the present invention, and constitutes explanation
A part of book, is used to explain the present invention together with the embodiment of the present invention, is not construed as limiting the invention.In the accompanying drawings,
Identical reference label typically represents same parts or step.
Fig. 1 is set for realizing the detection method of dog face characteristic point according to an embodiment of the present invention and the exemplary electron of device
Standby schematic block diagram;
Fig. 2 is the schematic flow chart of the detection method of dog face characteristic point according to an embodiment of the present invention;
Fig. 3 is the exemplary diagram of dog face characteristic point according to an embodiment of the present invention;
Fig. 4 is the schematic block diagram of the detection device of dog face characteristic point according to an embodiment of the present invention;
Fig. 5 is the schematic block diagram of the detection system of dog face characteristic point according to an embodiment of the present invention.
Specific embodiment
In order to enable the object, technical solutions and advantages of the present invention become apparent, root is described in detail below with reference to accompanying drawings
According to example embodiments of the present invention.Obviously, described embodiment is only a part of the embodiments of the present invention, rather than this hair
Bright whole embodiments, it should be appreciated that the present invention is not limited by example embodiment described herein.Based on described in the present invention
The embodiment of the present invention, those skilled in the art's obtained all other embodiment in the case where not making the creative labor
It should all fall under the scope of the present invention.
Firstly, being described with reference to Figure 1 the detection method and device of dog face characteristic point for realizing the embodiment of the present invention
Exemplary electronic device 100.
As shown in Figure 1, electronic equipment 100 include one or more processors 101, it is one or more storage device 102, defeated
Enter device 103, output device 104, imaging sensor 105, the company that these components pass through bus system 106 and/or other forms
The interconnection of connection mechanism (not shown).It should be noted that the component and structure of electronic equipment 100 shown in FIG. 1 are only exemplary, rather than
Restrictive, as needed, the electronic equipment also can have other assemblies and structure.
The processor 101 can be central processing unit (CPU) or have data-handling capacity and/or instruction execution
The processing unit of the other forms of ability, and the other components that can control in the electronic equipment 100 are desired to execute
Function.
The storage device 102 may include one or more computer program products, and the computer program product can
To include various forms of computer readable storage mediums, such as volatile memory and/or nonvolatile memory.It is described easy
The property lost memory for example may include random access memory (RAM) and/or cache memory (cache) etc..It is described non-
Volatile memory for example may include read-only memory (ROM), hard disk, flash memory etc..In the computer readable storage medium
On can store one or more computer program instructions, processor 102 can run described program instruction, to realize hereafter institute
The client functionality (realized by processor) in the embodiment of the present invention stated and/or other desired functions.In the meter
Can also store various application programs and various data in calculation machine readable storage medium storing program for executing, for example, the application program use and/or
The various data etc. generated.
The input unit 103 can be the device that user is used to input instruction, and may include keyboard, mouse, wheat
One or more of gram wind and touch screen etc..
The output device 104 can export various information (such as image or sound) to external (such as user), and
It may include one or more of display, loudspeaker etc..
Described image sensor 105 can be shot the desired image of user (such as photo, video etc.), and will be captured
Image be stored in the storage device 102 for other components use.
Illustratively, for realizing the detection method of dog face characteristic point according to an embodiment of the present invention and the example of device electricity
Sub- equipment may be implemented as smart phone, tablet computer, video acquisition end of access control system etc..
The detection method 200 of dog face characteristic point according to an embodiment of the present invention is described next, with reference to Fig. 2.The method
200 include:
Characteristic point detection is carried out based on the image comprising dog face and trained detection model, obtains the feature of fine positioning
Point;
Wherein, the detection model includes first order network and second level network, the characteristic point packet for obtaining fine positioning
It includes:
The first order network of full-face images and detection model based on the dog face carries out characteristic point detection, obtains coarse positioning
Characteristic point;
The second level network of topography and detection model based on the dog face positions the characteristic point of coarse positioning,
Obtain the characteristic point of the fine positioning.
Wherein, the characteristic point of full-face images rough estimate dog face of the first order network based on dog face, obtains the dog face
The characteristic point of coarse positioning;In order to further increase the precision of characteristic point detection, base of the second level network in the characteristic point of coarse positioning
On plinth, the topography based on dog face is adjusted the characteristic point of coarse positioning, finally obtains pinpoint characteristic point.And base
In the first order network that the full face information of dog face is established, and the cascade mind of the second level network composition based on local message foundation
Through network, Query refinement predicts the location information of dog face characteristic point, realizes the high accuracy positioning of dog face characteristic point.Illustratively,
The detection method of dog face characteristic point according to an embodiment of the present invention can with memory and processor unit or
It is realized in system.
The detection method of dog face characteristic point according to an embodiment of the present invention can be deployed at Image Acquisition end, for example, can
To be deployed in the Image Acquisition end of access control system;It can be deployed at personal terminal, such as smart phone, tablet computer, individual
Computer etc..Alternatively, the detection method of dog face characteristic point according to an embodiment of the present invention can also be deployed in service with being distributed
At device end (or cloud) and personal terminal.
The detection method of dog face characteristic point according to an embodiment of the present invention passes through the grade based on full face and local message foundation
Join neural network, Query refinement predicts the location information of dog face characteristic point, realizes the high accuracy positioning of dog face characteristic point, can
The accuracy and real-time for effectively improving the detection of dog face characteristic point can be widely applied to about the various of dog face image processing
Occasion.
According to embodiments of the present invention, the method 200 further include: based on pre-defined rule to training sample image full-face images
Characteristic point mark is carried out with the dog face in training sample topography;
Based on after mark training sample full-face images and training sample topography be trained, obtain trained inspection
Survey model.
Illustratively, the pre-defined rule includes ear's profile, eye profile, nasal contours, mouth based on the dog face
In contouring, face mask at least one in carry out characteristic point mark.
Illustratively, it includes: the right boundary for marking the basal part of the ear that ear's profile based on the dog face, which carries out characteristic point mark,
It characteristic point, basal part of the ear central feature point, have sharp ears characteristic point and is equidistantly marked on the basis of basal part of the ear central feature point to have sharp ears characteristic point
The characteristic point of note.
Illustratively, it includes: mark left eye central feature point, in left eye that the eye profile based on the dog face, which is labeled,
The characteristic point and left eye central vertical line and the eye profile that edema with the heart involved horizontal line intersects with the left and right sides of the left eye profile
Two sides up and down intersection characteristic point;With
The feature that mark right eye central feature point, right eye central horizontal line intersect with the left and right sides of the right eye profile
The characteristic point that point and right eye central vertical line intersect with the two sides up and down of the right eye profile.
Illustratively, the eye profile based on the dog face is labeled further include: with the left eye central horizontal line and
On the basis of the characteristic point that left eye central vertical line intersects with the left eye profile, characteristic point is equidistantly marked along the left eye profile;
With
On the basis of the characteristic point that the right eye central horizontal line and right eye central vertical line intersect the right eye profile,
Characteristic point is equidistantly marked along the right eye profile.
Illustratively, it includes: mark nose central feature point that the nasal contours based on the dog face, which are labeled,.
Illustratively, it includes: mark left side corners of the mouth characteristic point, upper lip that the mouth profile based on the dog face, which is labeled,
Left side profile corner feature point, upper lip central feature point, upper lip right lateral contours corner feature point, right side corners of the mouth characteristic point,
Profile corner feature point, lower lip central feature point, lower lip right lateral contours corner feature point on the left of lower lip.
In one embodiment, it includes: to trigger from the side corners of the mouth that the mouth profile based on the dog face, which is labeled, along
Upper lip successively marks the side corners of the mouth characteristic point, the upper lip side profile corner feature point, upper lip central feature point, upper lip
Another side profile corner feature point, other side corners of the mouth characteristic point;
Successively mark the lower lip side profile corner feature point along lower lip, lower lip central feature point, lower lip is another
Side profile corner feature point.
Illustratively, the face mask based on the dog face, which is labeled, includes:
Edema with the heart involved in characteristic point that mark crown central feature point, left eye central horizontal line and face left side profile intersect, nose
The characteristic point and nose of characteristic point, right eye central horizontal line and the intersection of face right lateral contours that horizontal line intersects with profile on the left of face
The characteristic point that central horizontal line intersects with face right lateral contours.
Illustratively, the face mask based on the dog face is labeled further include:
The characteristic point and the nose central horizontal line and face intersected with the left eye central horizontal line with profile on the left of face
Point on the basis of the characteristic point of left side profile intersection equidistantly marks characteristic point along profile on the left of face (as from top to bottom);With described
The feature for characteristic point and the nose central horizontal line and face the right lateral contours intersection that right eye center line intersects with face right lateral contours
Point on the basis of point, equidistantly marks characteristic point along face right lateral contours.
In one embodiment, as shown in figure 3, Fig. 3 shows the characteristic point according to an embodiment of the present invention comprising dog face
Example images.Referring to Fig. 3, further illustrates by taking a dog face image as an example and dog face is labeled according to pre-defined rule.Tool
For body, chosen properly based on five ear of dog face, face, nose, mouth, face mask main parts in dog face image shown in Fig. 3
Characteristic point be labeled.
Firstly, ear's profile based on the dog face carries out the mark of characteristic point;It specifically includes: ear's wheel based on dog face
Exterior feature marks the right boundary characteristic point 1,2 of the left ear basal part of the ear, can mark two characteristic points 1,2 from top to bottom from the left ear basal part of the ear;Mark
Infuse left ear basal part of the ear central feature point 3, left ear have sharp ears characteristic point 4, with left ear basal part of the ear central feature point 3 to left ear have sharp ears characteristic point 4
On the basis of the characteristic point 5,6 that equidistantly marks;Likewise, the right boundary characteristic point 7,8 of the mark auris dextra basal part of the ear, it can be from auris dextra ear
Root marks two characteristic points 7,8 from top to bottom;Mark auris dextra basal part of the ear central feature point 9, auris dextra have sharp ears characteristic point 10, with auris dextra ear
The characteristic point 11,12 equidistantly marked on the basis of root central feature point 9 to auris dextra have sharp ears characteristic point 10.
Then, the eye profile based on the dog face is labeled;It specifically includes: mark left eye central feature point 14, a left side
Characteristic point 15,16 that eye central horizontal line intersects with the left eye profile left and right sides and left eye central vertical line with it is described
The left eye profile characteristic point 17,18 that two sides are intersected up and down;With the left eye central horizontal line and left eye central vertical line with it is described
On the basis of 15,16,17,18 4 characteristic points of characteristic point of left eye profile intersection, left eye profile is divided into four bulks, from upper left side
It sets out clockwise, every piece equidistantly marks a characteristic point, marks 4 characteristic points altogether: 19,20,21,22;
Characteristic point 24 that mark right eye central feature point 23, right eye central horizontal line intersect with the right eye profile left and right sides,
25, the intersecting features point 26,27 of right eye central vertical line and two sides above and below right eye profile;With the right eye central horizontal line and the right side
On the basis of 24,25,26,27 4 characteristic points of characteristic point that eye central vertical line intersects the right eye profile, by right eye profile
29 it is divided into four bulks, from upper left side clockwise, every piece equidistantly marks a characteristic point, 4 characteristic points are marked altogether: 28,
30、31。
Then, the nasal contours based on the dog face are labeled, and are specifically included: mark nose central feature point 32.
Then, the face mask based on the dog face is labeled, and is specifically included: mark crown central feature point 13, a left side
The characteristic point 33 that eye central horizontal line intersects with profile on the left of face, the characteristic point that nose central horizontal line intersects with profile on the left of face
34;On the basis of above-mentioned 33,34 two characteristic points, characteristic point is equidistantly marked along profile on the left of face, can be from top to bottom etc.
Away from marking two characteristic points 35,36;
The characteristic point 37 that mark right eye central horizontal line intersects with face right lateral contours, nose central horizontal line and face right-hand wheel
The characteristic point 38 of exterior feature intersection;On the basis of above-mentioned 37,38 two characteristic points, characteristic point is equidistantly marked along face right lateral contours, it can
To be equidistantly to mark two characteristic points 39,40 from top to bottom.
Finally, the mouth profile based on the dog face is labeled, specifically include: mark left side corners of the mouth characteristic point 41, on
Profile corner feature point 42, upper lip central feature point 43, upper lip right lateral contours corner feature point 44, right side mouth on the left of lip
Profile corner feature point 46, lower lip central feature point 47, lower lip right lateral contours inflection point are special on the left of corner characteristics point 45, lower lip
Sign point 48;
It can be from the left side corners of the mouth, it is left successively to mark left side corners of the mouth characteristic point 41, upper lip along upper lip profile
Edge profile corner feature point 42, upper lip central feature point 43, upper lip right edge outline corner feature point 44, the right corners of the mouth feature
Point 45;Lower lip left side profile corner feature point 46, lower lip central feature point 47, lower mouth are successively marked along lower lip profile
Lip right edge outline corner feature point 48.It is appreciated that can also based on from the right corners of the mouth, along upper lip profile and under
Lip outline marks corresponding characteristic point.
It follows that in the present embodiment, characteristic point mark is carried out to dog face based on pre-defined rule, finally for every ear
Piece it is labelled with 6 characteristic points, each eyes are labelled with 9 characteristic points, lip position is labelled with 8 characteristic points, noses and is labelled with 1
A characteristic point, face profile be labelled with 9 characteristic points, 48 characteristic points altogether.
It should be noted that above-mentioned annotation step is only example, the pre-defined rule is not represented;The pre-defined rule
For mark sequence, there is no limit.In addition, the pre-defined rule can increase characteristic point with actual conditions according to the design needs
Number provides good data basis to improve the accuracy of mark for down-stream.
According to embodiments of the present invention, the method 200 further include:
The full face sample image of dog face before being marked based on the training sample image before mark, and based on mark
The training sample image afterwards marked after the full face sample image of dog face;
First nerves network is trained according to the dog face full face sample image after mark, obtains trained detection mould
The first order network of type.
According to embodiments of the present invention, the method 200 further include: according to the position of dog face to the full face figure of the dog face
As being split, the topography of the dog face is obtained.
Wherein, if topography is that dog face full-face images are divided into stem portion according to dog face, each part is wrapped
Image containing complete dog face.Such as, dog face includes: left ear, auris dextra, left eye, right eye, nose, mouth, left face, right face
Deng then the topography of dog face can be the image including left ear, auris dextra, left eye, right eye, nose, mouth, left face or right face.
According to embodiments of the present invention, the method 200 further include:
Dog face fractional sample image before being marked based on the training sample image before mark, and based on mark
The training sample image afterwards marked after dog face fractional sample image;
Nervus opticus network is trained according to the dog face fractional sample image after mark, obtains trained detection mould
The second level network of type.
According to embodiments of the present invention, the method 200 further comprises: topography and detection mould based on the dog face
The second level network of type positions the characteristic point of coarse positioning, and the characteristic point for obtaining the fine positioning includes:
The second level network of topography and detection model based on the dog face positions the characteristic point of coarse positioning,
Obtain the characteristic point of topography;
The characteristic point of topography is coordinately transformed and is integrated, the characteristic point of the fine positioning of the dog face is obtained.
Because the first order network of detection model carries out characteristic point detection to the full-face images of the dog face, obtain complete
The coarse positioning characteristic point of face image;It is special to the coarse positioning by second level network in order to further increase the precision of characteristic point
Sign clicks through a successive step of advancing.Since second level network is the topography based on dog face, characteristic point is carried out using local message
The topography of detection, input second level network should be that the corresponding same topography of homolog is normalized to standard
Size, and second level network is inputted after rotation into alignment to unified angle, the output of first order network can be divided into it is each
The characteristic point of the corresponding multiple groups coarse positioning of topography, each topography are thick to corresponding one group by second level network
The characteristic point of positioning is adjusted, and obtains one group of characteristic point corresponding with each topography, by each topography pair
The characteristic point answered is coordinately transformed and integrates the characteristic point of the fine positioning of available entire dog face.
Illustratively, the characteristic point of topography is coordinately transformed and is integrated to obtain the spy of the fine positioning of the dog face
Sign point, comprising:
Obtain reference position and rotation angle of the topography of dog face relative to full-face images;
The characteristic point of corresponding topography is coordinately transformed according to the reference position and rotation angle, is obtained
The characteristic point of transformed topography;
The characteristic point of topography after each transformation is integrated, the characteristic point of the fine positioning of dog face is obtained.
Wherein, the coordinate of local image characteristics point is to determine that therefore, it is necessary to obtain part relative to itself topography
Image obtains the position in full-face images of local image characteristics point according to reference position with respect to the reference position of full-face images
It sets.In addition, therefore, it is necessary to according to the relatively full face of topography since the topography for being input to the second network is rotated
The rotation angle of image is by the characteristic point rotation transformation of topography to the position in full face.The coordinate transform is from two dimension
Coordinate to two-dimensional coordinate linear transformation, and keep image grazing and collimation, in plane object pose change,
Input picture and transformation matrix can be subjected to matrix multiplication and obtain transformed image coordinate by matrix description.
During in the second level, network carries out characteristic point detection to topography, by topography with respect to full-face images
Reference position carries out rotation into alignment to the same angle, and then ensure that the accuracy of characteristic point detection.
According to embodiments of the present invention, the method 200 further include: output includes the dog face image of the dog face characteristic point.
Fig. 4 shows the schematic block diagram of the detection device 300 of dog face characteristic point according to an embodiment of the present invention.Such as Fig. 4
Shown, the detection device 400 of dog face characteristic point according to an embodiment of the present invention includes:
Detection module 410 is obtained for carrying out characteristic point detection based on the image comprising dog face and trained detection model
To the characteristic point of fine positioning;
Wherein, the detection model includes first order network 411 and second level network 412, and the first order network 411 is used
Characteristic point detection is carried out in the full-face images to the dog face, obtains the characteristic point of coarse positioning;
The second level network 412 is used to position the characteristic point of the coarse positioning according to the topography of dog face,
Obtain the characteristic point of the fine positioning.
According to embodiments of the present invention, the training of the detection model includes: complete to training sample image based on pre-defined rule
Dog face in face image and training sample topography carries out characteristic point mark;
Based on after mark training sample full-face images and training sample topography detection model is trained, obtain
Trained detection model.
Illustratively, the pre-defined rule includes ear's profile, eye profile, nasal contours, mouth based on the dog face
In contouring, face mask at least one in carry out characteristic point mark.
Illustratively, it includes: the left and right side for marking the basal part of the ear that ear's profile based on the dog face, which carries out the mark of characteristic point,
Boundary's characteristic point, basal part of the ear central feature point, have sharp ears characteristic point and on the basis of basal part of the ear central feature point to have sharp ears characteristic point it is equidistant
The characteristic point of mark.
Illustratively, it includes: mark left eye central feature that the eye profile based on the dog face, which carries out the mark of characteristic point,
Point, the characteristic point and left eye central vertical line and institute that left eye central horizontal line intersects with the left and right sides of the left eye profile
State the characteristic point of the intersection of two sides up and down of eye profile;With
The feature that mark right eye central feature point, right eye central horizontal line intersect with the left and right sides of the right eye profile
The characteristic point that point and right eye central vertical line intersect with the two sides up and down of the right eye profile.
Illustratively, the eye profile based on the dog face carries out the mark of characteristic point further include: with the left eye center
On the basis of the characteristic point that horizontal line and left eye central vertical line intersect with the left eye profile, equidistantly marked along the left eye profile
Characteristic point;With
On the basis of the characteristic point that the right eye central horizontal line and right eye central vertical line intersect the right eye profile,
Characteristic point is equidistantly marked along the right eye profile.
Illustratively, it includes: mark nose central feature that the nasal contours based on the dog face, which carry out the mark of characteristic point,
Point.
Illustratively, it includes: mark left side corners of the mouth feature that the mouth profile based on the dog face, which carries out the mark of characteristic point,
Profile corner feature point, upper lip central feature point, upper lip right lateral contours corner feature point, right side mouth on the left of point, upper lip
Profile corner feature point, lower lip central feature point, lower lip right lateral contours corner feature point on the left of corner characteristics point, lower lip.
Illustratively, the mark of the face mask progress characteristic point based on the dog face includes:
Mark crown central feature point, the characteristic point that left eye central horizontal line intersects with profile on the left of face, edema with the heart involved in nose
The characteristic point and nose of characteristic point, right eye central horizontal line and the intersection of face right lateral contours that horizontal line intersects with profile on the left of face
The characteristic point that central horizontal line intersects with face right lateral contours.
Illustratively, the face mask based on the dog face carries out the mark of characteristic point further include:
The characteristic point and the nose central horizontal line and face intersected with the left eye central horizontal line with profile on the left of face
Point on the basis of the characteristic point of left side profile intersection equidistantly marks characteristic point along profile on the left of face;With the right eye center line with
Point on the basis of the characteristic point that the characteristic point of face right lateral contours intersection and the nose central horizontal line intersect with face right lateral contours, edge
Face right lateral contours equidistantly mark characteristic point.
According to embodiments of the present invention, the training of the first order network of the detection model includes:
The full face sample image of dog face before being marked based on the training sample image before mark, and based on mark
The training sample image afterwards marked after the full face sample image of dog face;
First nerves network is trained according to the dog face full face sample image after mark, obtains trained detection mould
The first order network of type.
According to embodiments of the present invention, the detection module 410 is also used to: complete to the dog face according to the position of dog face
Face image is split, and obtains the topography of the dog face.
According to embodiments of the present invention, the training of the second level network of the detection model includes:
Dog face fractional sample image before being marked based on the training sample image before mark, and based on mark
The training sample image afterwards marked after dog face fractional sample image;
Nervus opticus network is trained according to the dog face fractional sample image after mark, obtains trained detection mould
The second level network of type.
According to embodiments of the present invention, the second level network 412 is further used for:
Topography based on the dog face positions the characteristic point of coarse positioning, obtains the characteristic point of topography;
The characteristic point of topography is coordinately transformed and is integrated, the characteristic point of the fine positioning of the dog face is obtained.
Illustratively, the second level network 412 is further used for:
Obtain reference position and rotation angle of the topography of dog face relative to full-face images;
The characteristic point of corresponding topography is coordinately transformed according to the reference position and rotation angle, is obtained
The characteristic point of transformed topography;
The characteristic point of topography after each transformation is integrated, the characteristic point of the fine positioning of dog face is obtained.
According to embodiments of the present invention, described device 400 further include: output module 420 includes that the dog face is special for exporting
The dog face image of sign point and/or the dog face characteristic point coordinate.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
Fig. 5 shows the schematic block diagram of the mark detection 500 of dog face characteristic point according to an embodiment of the present invention.Dog face is special
The detection system 500 of sign point includes imaging sensor 510, storage device 530 and processor 540.
Imaging sensor 510 is for acquiring image data.
In detection method of the storage of storage device 530 for realizing dog face characteristic point according to an embodiment of the present invention
The program code of corresponding steps.
The processor 540 is for running the program code stored in the storage device 530, to execute according to the present invention
The corresponding steps of the detection method of the dog face characteristic point of embodiment, and for realizing dog face feature according to an embodiment of the present invention
Detection module 410 in the detection device of point.
In addition, according to embodiments of the present invention, additionally providing a kind of storage medium, storing program on said storage
Instruction, when described program instruction is run by computer or processor for executing the mark of the dog face characteristic point of the embodiment of the present invention
The corresponding steps of injecting method, and for realizing the respective mode in the annotation equipment of dog face characteristic point according to an embodiment of the present invention
Block.The storage medium for example may include the storage card of smart phone, the storage unit of tablet computer, personal computer it is hard
Disk, read-only memory (ROM), Erasable Programmable Read Only Memory EPROM (EPROM), portable compact disc read-only memory (CD-
ROM), any combination of USB storage or above-mentioned storage medium.The computer readable storage medium can be one or
Any combination of multiple computer readable storage mediums, such as a computer readable storage medium include for being randomly generated
The computer-readable program code of action command sequence, another computer readable storage medium include special for carrying out dog face
Levy the computer-readable program code of the mark of point.
In one embodiment, the computer program instructions may be implemented real according to the present invention when being run by computer
Each functional module of the detection device of the dog face characteristic point of example is applied, and/or can be executed according to an embodiment of the present invention
The detection method of dog face characteristic point.
Each module in the detection system of dog face characteristic point according to an embodiment of the present invention can be by real according to the present invention
The processor for applying the electronic equipment of the detection of the dog face characteristic point of example, which runs the computer program instructions stored in memory, to be come
It realizes, or the meter that can be stored in the computer readable storage medium of computer program product according to an embodiment of the present invention
The realization when instruction of calculation machine is run by computer.
Detection method, device, system and the storage medium of dog face characteristic point according to an embodiment of the present invention, by being based on
The cascade neural network that full face and local message are established, Query refinement predict the location information of dog face characteristic point, realize dog face
The high accuracy positioning of characteristic point can effectively improve the accuracy and real-time of the detection of dog face characteristic point, can be widely applied
In the various occasions handled about dog face image.
Although describing example embodiment by reference to attached drawing here, it should be understood that above example embodiment are only exemplary
, and be not intended to limit the scope of the invention to this.Those of ordinary skill in the art can carry out various changes wherein
And modification, it is made without departing from the scope of the present invention and spiritual.All such changes and modifications are intended to be included in appended claims
Within required the scope of the present invention.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
In several embodiments provided herein, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, apparatus embodiments described above are merely indicative, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can be tied
Another equipment is closed or is desirably integrated into, or some features can be ignored or not executed.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the present invention and help to understand one or more of the various inventive aspects,
To in the description of exemplary embodiment of the present invention, each feature of the invention be grouped together into sometimes single embodiment, figure,
Or in descriptions thereof.However, the method for the invention should not be construed to reflect an intention that i.e. claimed
The present invention claims features more more than feature expressly recited in each claim.More precisely, as corresponding
As claims reflect, inventive point is that all features less than some disclosed single embodiment can be used
Feature solves corresponding technical problem.Therefore, it then follows thus claims of specific embodiment are expressly incorporated in the tool
Body embodiment, wherein each, the claims themselves are regarded as separate embodiments of the invention.
It will be understood to those skilled in the art that any combination pair can be used other than mutually exclusive between feature
All features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed any method
Or all process or units of equipment are combined.Unless expressly stated otherwise, this specification (is wanted including adjoint right
Ask, make a summary and attached drawing) disclosed in each feature can be replaced with an alternative feature that provides the same, equivalent, or similar purpose.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments
In included certain features rather than other feature, but the combination of the feature of different embodiments mean it is of the invention
Within the scope of and form different embodiments.For example, in detail in the claims, embodiment claimed it is one of any
Can in any combination mode come using.
Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors
Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice
Microprocessor or digital signal processor (DSP) realize some moulds in article analytical equipment according to an embodiment of the present invention
The some or all functions of block.The present invention is also implemented as a part or complete for executing method as described herein
The program of device (for example, computer program and computer program product) in portion.It is such to realize that program of the invention can store
On a computer-readable medium, it or may be in the form of one or more signals.Such signal can be from internet
Downloading obtains on website, is perhaps provided on the carrier signal or is provided in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability
Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real
It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch
To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame
Claim.
The above description is merely a specific embodiment or to the explanation of specific embodiment, protection of the invention
Range is not limited thereto, and anyone skilled in the art in the technical scope disclosed by the present invention, can be easily
Expect change or replacement, should be covered by the protection scope of the present invention.Protection scope of the present invention should be with claim
Subject to protection scope.
Claims (16)
1. a kind of detection method of dog face characteristic point, which is characterized in that the described method includes:
Characteristic point detection is carried out based on the image comprising dog face and trained detection model, obtains the characteristic point of fine positioning;
Wherein, the detection model includes first order network and second level network, and the characteristic point for obtaining fine positioning includes:
The first order network of full-face images and detection model based on the dog face carries out characteristic point detection, obtains the spy of coarse positioning
Sign point;
The second level network of topography and detection model based on the dog face positions the characteristic point of coarse positioning, obtains
The characteristic point of the fine positioning.
2. detection method as described in claim 1, which is characterized in that the method also includes: according to the position of dog face
The dog face full-face images are split, the topography of the dog face is obtained.
3. detection method as claimed in claim 2, which is characterized in that topography and detection model based on the dog face
Second level network positions the characteristic point of coarse positioning, and the characteristic point for obtaining the fine positioning includes:
The second level network of topography and detection model based on the dog face positions the characteristic point of coarse positioning, obtains
The characteristic point of topography;
The characteristic point of topography is coordinately transformed and is integrated to obtain the characteristic point of the fine positioning of the dog face.
4. detection method as claimed in claim 3, which is characterized in that by the characteristic point of topography be coordinately transformed with it is whole
It closes and obtains the characteristic point of the fine positioning of the dog face, comprising:
Obtain reference position and rotation angle of the topography of dog face relative to full-face images;
The characteristic point of corresponding topography is coordinately transformed according to the reference position and rotation angle, is converted
The characteristic point of topography afterwards;
The characteristic point of topography after each transformation is integrated, the characteristic point of the fine positioning of dog face is obtained.
5. detection method as described in claim 1, which is characterized in that the training of the detection model includes: based on pre- set pattern
Characteristic point mark then is carried out to the dog face in training sample full-face images and training sample topography;
Based on after mark training sample full-face images and training sample topography detection model is trained, trained
Good detection model.
6. detection method as claimed in claim 5, which is characterized in that the pre-defined rule includes the ear based on the dog face
The mark of at least one of profile, eye profile, nasal contours, mouth profile, face mask progress characteristic point.
7. detection method as claimed in claim 6, which is characterized in that ear's profile based on the dog face carries out characteristic point
Mark includes: to mark the right boundary characteristic point of the basal part of the ear, basal part of the ear central feature point, have sharp ears characteristic point and with basal part of the ear center spy
The characteristic point equidistantly marked on the basis of sign point to have sharp ears characteristic point.
8. detection method as claimed in claim 6, which is characterized in that the eye profile based on the dog face carries out characteristic point
Mark includes: the feature for marking left eye central feature point, left eye central horizontal line and intersecting with the left and right sides of the left eye profile
The characteristic point that point and left eye central vertical line intersect with the two sides up and down of the left eye profile;With
Characteristic point that mark right eye central feature point, right eye central horizontal line intersect with the left and right sides of the right eye profile, with
And the characteristic point that right eye central vertical line intersects with the two sides up and down of the right eye profile.
9. detection method as claimed in claim 8, which is characterized in that the eye profile based on the dog face carries out characteristic point
Mark further include: the characteristic point intersected using the left eye central horizontal line and left eye central vertical line with the left eye profile is base
Standard equidistantly marks characteristic point along the left eye profile;With
On the basis of the characteristic point that the right eye central horizontal line and right eye central vertical line intersect the right eye profile, along institute
It states right eye profile and equidistantly marks characteristic point.
10. detection method as claimed in claim 6, which is characterized in that the nasal contours based on the dog face are labeled packet
It includes: mark nose central feature point.
11. detection method as claimed in claim 6, which is characterized in that the mouth profile based on the dog face is labeled packet
It includes: profile corner feature point, upper lip central feature point, upper lip right-hand wheel on the left of mark left side corners of the mouth characteristic point, upper lip
Wide corner feature point, right side corners of the mouth characteristic point, lower lip left side profile corner feature point, lower lip central feature point, lower lip
Right lateral contours corner feature point.
12. detection method as claimed in claim 6, which is characterized in that the face mask based on the dog face is labeled packet
It includes:
Mark crown central feature point, the characteristic point that left eye central horizontal line intersects with profile on the left of face, nose central horizontal line
The characteristic point of the characteristic point, right eye central horizontal line and the intersection of face right lateral contours that intersect with profile on the left of face and nose center
The characteristic point that horizontal line intersects with face right lateral contours.
13. detection method as claimed in claim 12, which is characterized in that the face mask based on the dog face is labeled also
It include: characteristic point intersect with profile on the left of the left eye central horizontal line and face and the nose central horizontal line and a face left side
On the basis of the characteristic point of side profile intersection, characteristic point is equidistantly marked along profile on the left of face;It is right with the right eye center line and face
Point on the basis of the characteristic point that the characteristic point of side profile intersection and the nose central horizontal line intersect with face right lateral contours, along
Face right lateral contours equidistantly mark characteristic point.
14. a kind of detection device of dog face characteristic point, which is characterized in that described device includes:
It is fixed to obtain essence for carrying out characteristic point detection based on the image comprising dog face and trained detection model for detection module
The characteristic point of position;
Wherein, the detection model includes first order network and second level network, and the first order network is used for the dog face
Full-face images carry out characteristic point detection, obtain the characteristic point of coarse positioning;
The second level network obtains described for being positioned according to the topography of dog face to the characteristic point of the coarse positioning
The characteristic point of fine positioning.
15. a kind of detection system of dog face characteristic point, including memory, processor and it is stored on the memory and described
The computer program run on processor, which is characterized in that realize that right is wanted when the processor executes the computer program
The step of seeking any one of 1 to 13 the method.
16. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of any one of claims 1 to 13 the method is realized when being computer-executed.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010327006.5A CN111695405B (en) | 2018-12-28 | 2018-12-28 | Dog face feature point detection method, device and system and storage medium |
CN201811628345.6A CN109829380B (en) | 2018-12-28 | 2018-12-28 | Method, device and system for detecting dog face characteristic points and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811628345.6A CN109829380B (en) | 2018-12-28 | 2018-12-28 | Method, device and system for detecting dog face characteristic points and storage medium |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010327006.5A Division CN111695405B (en) | 2018-12-28 | 2018-12-28 | Dog face feature point detection method, device and system and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109829380A true CN109829380A (en) | 2019-05-31 |
CN109829380B CN109829380B (en) | 2020-06-02 |
Family
ID=66861476
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811628345.6A Active CN109829380B (en) | 2018-12-28 | 2018-12-28 | Method, device and system for detecting dog face characteristic points and storage medium |
CN202010327006.5A Active CN111695405B (en) | 2018-12-28 | 2018-12-28 | Dog face feature point detection method, device and system and storage medium |
Family Applications After (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010327006.5A Active CN111695405B (en) | 2018-12-28 | 2018-12-28 | Dog face feature point detection method, device and system and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (2) | CN109829380B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110909618A (en) * | 2019-10-29 | 2020-03-24 | 泰康保险集团股份有限公司 | Pet identity recognition method and device |
CN114155240A (en) * | 2021-12-13 | 2022-03-08 | 韩松洁 | Ear acupoint detection method and device and electronic equipment |
CN115240230A (en) * | 2022-09-19 | 2022-10-25 | 星宠王国(北京)科技有限公司 | Canine face detection model training method and device, and detection method and device |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113723332A (en) * | 2021-09-07 | 2021-11-30 | 中国工商银行股份有限公司 | Facial image recognition method and device |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070071289A1 (en) * | 2005-09-29 | 2007-03-29 | Kabushiki Kaisha Toshiba | Feature point detection apparatus and method |
CN103208133A (en) * | 2013-04-02 | 2013-07-17 | 浙江大学 | Method for adjusting face plumpness in image |
CN103971112A (en) * | 2013-02-05 | 2014-08-06 | 腾讯科技(深圳)有限公司 | Image feature extracting method and device |
CN104951743A (en) * | 2015-03-04 | 2015-09-30 | 苏州大学 | Active-shape-model-algorithm-based method for analyzing face expression |
CN107403141A (en) * | 2017-07-05 | 2017-11-28 | 中国科学院自动化研究所 | Method for detecting human face and device, computer-readable recording medium, equipment |
CN107704817A (en) * | 2017-09-28 | 2018-02-16 | 成都品果科技有限公司 | A kind of detection algorithm of animal face key point |
CN108875480A (en) * | 2017-08-15 | 2018-11-23 | 北京旷视科技有限公司 | A kind of method for tracing of face characteristic information, apparatus and system |
CN108985210A (en) * | 2018-07-06 | 2018-12-11 | 常州大学 | A kind of Eye-controlling focus method and system based on human eye geometrical characteristic |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103295024B (en) * | 2012-02-29 | 2017-03-01 | 佳能株式会社 | Classification and method for checking object and device and image taking and processing equipment |
CN103218610B (en) * | 2013-04-28 | 2016-06-08 | 宁波江丰生物信息技术有限公司 | The forming method of dog face detector and dog face detecting method |
CN103824049A (en) * | 2014-02-17 | 2014-05-28 | 北京旷视科技有限公司 | Cascaded neural network-based face key point detection method |
-
2018
- 2018-12-28 CN CN201811628345.6A patent/CN109829380B/en active Active
- 2018-12-28 CN CN202010327006.5A patent/CN111695405B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070071289A1 (en) * | 2005-09-29 | 2007-03-29 | Kabushiki Kaisha Toshiba | Feature point detection apparatus and method |
CN103971112A (en) * | 2013-02-05 | 2014-08-06 | 腾讯科技(深圳)有限公司 | Image feature extracting method and device |
CN103208133A (en) * | 2013-04-02 | 2013-07-17 | 浙江大学 | Method for adjusting face plumpness in image |
CN104951743A (en) * | 2015-03-04 | 2015-09-30 | 苏州大学 | Active-shape-model-algorithm-based method for analyzing face expression |
CN107403141A (en) * | 2017-07-05 | 2017-11-28 | 中国科学院自动化研究所 | Method for detecting human face and device, computer-readable recording medium, equipment |
CN108875480A (en) * | 2017-08-15 | 2018-11-23 | 北京旷视科技有限公司 | A kind of method for tracing of face characteristic information, apparatus and system |
CN107704817A (en) * | 2017-09-28 | 2018-02-16 | 成都品果科技有限公司 | A kind of detection algorithm of animal face key point |
CN108985210A (en) * | 2018-07-06 | 2018-12-11 | 常州大学 | A kind of Eye-controlling focus method and system based on human eye geometrical characteristic |
Non-Patent Citations (1)
Title |
---|
陈锐 等: "基于级联卷积神经网络的人脸关键点定位", 《四川理工学院学报( 自然科学版)》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110909618A (en) * | 2019-10-29 | 2020-03-24 | 泰康保险集团股份有限公司 | Pet identity recognition method and device |
CN110909618B (en) * | 2019-10-29 | 2023-04-21 | 泰康保险集团股份有限公司 | Method and device for identifying identity of pet |
CN114155240A (en) * | 2021-12-13 | 2022-03-08 | 韩松洁 | Ear acupoint detection method and device and electronic equipment |
CN115240230A (en) * | 2022-09-19 | 2022-10-25 | 星宠王国(北京)科技有限公司 | Canine face detection model training method and device, and detection method and device |
Also Published As
Publication number | Publication date |
---|---|
CN111695405B (en) | 2023-12-12 |
CN111695405A (en) | 2020-09-22 |
CN109829380B (en) | 2020-06-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109829380A (en) | A kind of detection method, device, system and the storage medium of dog face characteristic point | |
US11842438B2 (en) | Method and terminal device for determining occluded area of virtual object | |
CN106203305B (en) | Face living body detection method and device | |
CN109543663A (en) | A kind of dog personal identification method, device, system and storage medium | |
CN107798685B (en) | Pedestrian's height determines method, apparatus and system | |
CN106030661B (en) | The independent 3D scene texture background in the visual field | |
CN106650662B (en) | Target object shielding detection method and device | |
CN109146830A (en) | For generating the method, apparatus, system and storage medium of training data | |
US9747306B2 (en) | Method and apparatus for identifying input features for later recognition | |
CN108932456B (en) | Face recognition method, device and system and storage medium | |
CN108875524A (en) | Gaze estimation method, device, system and storage medium | |
CN108961149A (en) | Image processing method, device and system and storage medium | |
CN108875731A (en) | Target identification method, device, system and storage medium | |
CN106778453B (en) | Method and device for detecting glasses wearing in face image | |
CN106030610B (en) | The real-time 3D gesture recognition and tracking system of mobile device | |
CN108280455A (en) | Human body critical point detection method and apparatus, electronic equipment, program and medium | |
CN110008956A (en) | Invoice key message localization method, device, computer equipment and storage medium | |
US11113571B2 (en) | Target object position prediction and motion tracking | |
CN109325456A (en) | Target identification method, device, target identification equipment and storage medium | |
CN115661371B (en) | Three-dimensional object modeling method and device, computer equipment and storage medium | |
CN112766027A (en) | Image processing method, device, equipment and storage medium | |
CN104850823A (en) | Quality evaluation method and device for iris image | |
JP2022521540A (en) | Methods and systems for object tracking using online learning | |
CN110264523A (en) | A kind of method and apparatus of the location information of target image in determining test image | |
CN110490058A (en) | Training method, device, system and the computer-readable medium of pedestrian detection model |
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 |