CN113269170A - Intelligent portrait building block matching method and system based on feature similarity measurement - Google Patents
Intelligent portrait building block matching method and system based on feature similarity measurement Download PDFInfo
- Publication number
- CN113269170A CN113269170A CN202110816120.9A CN202110816120A CN113269170A CN 113269170 A CN113269170 A CN 113269170A CN 202110816120 A CN202110816120 A CN 202110816120A CN 113269170 A CN113269170 A CN 113269170A
- Authority
- CN
- China
- Prior art keywords
- building block
- feature
- image
- face
- matching
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
-
- 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
- G06V10/20—Image preprocessing
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
- G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
Abstract
The invention relates to a portrait building block intelligent matching method and a system based on feature similarity measurement. The measuring method based on the feature similarity can automatically match the building block image which is very similar to the input portrait, can realize the matching from the face picture to the portrait building block within the second-level time, is very efficient, has high personalization degree of the matched portrait building block, can well distinguish different face features, and really realizes personalized matching.
Description
Technical Field
The invention relates to the field of image processing, in particular to a portrait building block intelligent matching method and system based on feature similarity measurement.
Background
At present, personalized portrait building block design is usually designed manually by designers, namely after a user provides a portrait photo, the designers select parts from a template library to build and design according to some important characteristics of a face by using Studio, LDD and other design software, so as to design a portrait building block which is most similar to the face photo.
The manual design scheme has the defect that each portrait photo needs to be repeatedly designed, and a designer usually needs several days for the design process of a given portrait photo, so that the scheme is very inefficient in the block design of batch human face pictures, and cannot realize an automatic block design process.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides the intelligent matching method of the portrait building blocks based on the characteristic similarity measurement, the method can realize the matching from the portrait picture to the portrait building blocks in second-level time based on the deep learning algorithm and the characteristic similarity measurement, the personalized degree of the matched portrait building blocks is high, the different human face characteristics can be well distinguished, and the personalized matching is realized. The invention also relates to an intelligent portrait building block matching system based on the characteristic similarity measurement.
The technical scheme of the invention is as follows:
an intelligent portrait building block matching method based on feature similarity measurement is characterized by comprising the following steps:
acquiring a target image, and identifying the position of a face in the target image through a deep learning algorithm;
a characteristic extraction step: extracting a plurality of appearance features of the human face through a neural network algorithm;
and (3) feature matching: and respectively carrying out feature similarity measurement calculation on each extracted appearance feature of the human face and the corresponding building block image in a preset building block image library, and automatically matching the building block image with the highest similarity with the input human face.
Preferably, in the step of detecting and identifying the human face, when the human face in the target image cannot be identified through a deep learning algorithm, the operation is stopped; and if the target image contains a plurality of faces, respectively carrying out subsequent feature extraction and matching operation on each face.
Preferably, in the step of face detection and recognition, after the face position is recognized, the gender of the face is judged through a deep learning algorithm, and for the faces with different genders, the step of feature matching emphasizes on matching the faces to the building block images conforming to the corresponding gender features.
Preferably, in the feature extraction step, after the appearance features are extracted, the appearance features are further segmented, encoded and classified by a semantic segmentation algorithm.
Preferably, in the feature extraction step, the appearance features include five sense organs, a hair style, a hair color, a face shape, an expression and a skin color; for hair style characteristics, segmenting an input face image into hair style parts by a semantic segmentation algorithm, and further coding the characteristics of the hair style parts to measure hair styles of different types; for the hair color feature, the colors of the hair part are directly classified, and the hair color is identified.
Preferably, in the step of feature matching, corresponding feature values of five sense organs, hair style, hair color, face shape, expression and skin color are calculated for each image in the building block image library, feature similarity measurement calculation is performed on the feature values and the appearance features of the input face, and finally the cartoon building block image with the highest similarity to the input face is automatically matched.
Preferably, in the feature matching step, the feature similarity metric calculation includes a cosine similarity calculation and a pearson correlation coefficient calculation.
Preferably, the system comprises a face detection and recognition module, a feature extraction module and a feature matching module which are connected in sequence;
the face detection and recognition module acquires a target image and recognizes the position of a face in the target image through a deep learning algorithm;
the feature extraction module extracts a plurality of appearance features of the human face through a neural network algorithm;
and the feature matching module is used for respectively carrying out feature similarity measurement calculation on each extracted appearance feature of the human face and the corresponding building block image in the preset building block image library, and automatically matching the building block image with the highest similarity with the input human face.
Preferably, the appearance features extracted by the feature extraction module include five sense organs, hair style, hair color, face shape, expression and skin color; for hair style characteristics, segmenting an input face image into hair style parts by a semantic segmentation algorithm, and further coding the characteristics of the hair style parts to measure hair styles of different types; for the hair color feature, the colors of the hair part are directly classified, and the hair color is identified.
Preferably, the feature matching module calculates feature values of corresponding five sense organs, hair style, hair color, face shape, expression and skin color for each image in the building block image library, performs feature similarity measurement calculation with the appearance features of the input face, the feature similarity measurement calculation includes cosine similarity calculation and pearson correlation coefficient calculation, and finally automatically matches the cartoon building block image with the highest similarity with the input face.
The invention has the technical effects that:
the invention relates to a portrait building block intelligent matching method based on feature similarity measurement, which sequentially performs face detection and identification, feature extraction and feature matching, after a target image is obtained, the position of a face in the target image is identified through a deep learning algorithm, a plurality of appearance features of the face are extracted through a neural network algorithm, then each extracted appearance feature of the face is respectively subjected to feature similarity measurement calculation with a corresponding building block image in a preset building block image library, and finally a cartoon building block image with the highest similarity with the input face is automatically matched. The invention is essentially an intelligent matching method from a two-dimensional portrait picture to a three-dimensional portrait building block based on the measurement of feature similarity, based on a deep learning algorithm, a neural network algorithm and a measurement method of feature similarity, automatically extracts appearance features such as hairstyle, hair color, face shape, expression, skin color and the like in an input face picture through a neural network, then carries out feature similarity calculation and comparison according to the appearance features and building block images in a preset building block image library, can automatically match building block images which are very similar to the input portrait, does not need the manual design step of a designer in the traditional scheme, can realize matching from the portrait picture to the portrait building block within second-level time, is extremely high in efficiency, simultaneously, the personalization degree of the portrait building block matched by the method is high, and various appearance features such as hairstyle, hair color, expression and the like of the portrait are comprehensively considered, different human face characteristics can be well distinguished, and real personalized matching is realized.
The invention also relates to a human image building block intelligent matching system based on the characteristic similarity measurement, which corresponds to the human image building block intelligent matching method based on the characteristic similarity measurement and can be understood as a system for realizing the human image building block intelligent matching method based on the characteristic similarity measurement, the system mutually cooperates through a face detection and recognition module, a characteristic extraction module and a characteristic matching module which are sequentially executed, and after a target image is obtained by the face detection and recognition module, the position of a face in the target image is recognized through a depth learning algorithm; the feature extraction module extracts a plurality of appearance features of the face through a neural network algorithm; and the characteristic matching module calculates and compares the similarity of the characteristics according to the appearance characteristics and the building block images in a preset building block image library, and finally matches the building block image with the highest similarity with the input human face. The system can automatically extract various characteristics of the human face through the neural network model and carry out characteristic similarity calculation with images in the building block image library, so that cartoon building block images which are similar to each other can be automatically matched from the input human face.
Drawings
FIG. 1 is a flow chart of the intelligent portrait building block matching method based on feature similarity measurement according to the present invention.
FIG. 2 is a preferred flowchart of the intelligent matching method for portrait blocks based on feature similarity measurement according to the present invention.
FIG. 3 is a block diagram of a human image building block intelligent matching system based on feature similarity measurement.
Detailed Description
For a clearer understanding of the contents of the present invention, reference will be made to the accompanying drawings and examples.
The invention relates to a portrait building block intelligent matching method based on feature similarity measurement, the flow of which is shown in figure 1, and the method comprises the following steps: a human face detection and identification step, wherein a target image is obtained, and the position of a human face in the target image is identified through a deep learning algorithm; a feature extraction step, namely extracting a plurality of appearance features of the human face through a neural network algorithm; and a feature matching step, wherein feature similarity measurement calculation is carried out on each extracted appearance feature of the human face and the corresponding building block image in a preset building block image library respectively, and the building block image with the highest similarity with the input human face is automatically matched. The method can match the building block image which is very similar to the input portrait automatically, can match the portrait picture to the portrait building block within second time, has high personalization degree of the matched portrait building block, can better distinguish different human face characteristics, and realizes personalized matching.
In particular, a preferred flow chart is shown in fig. 2. In the step of face detection and recognition, after a target image is obtained (or a given portrait photo), the position of a face in the target image is recognized through a deep learning algorithm, then the gender of the face is judged through the deep learning algorithm, for the faces with different genders, a certain emphasis is placed in the step of feature matching, and the faces are quickly matched to the building block image conforming to the corresponding gender features.
It should be noted that if the target image does not have a face, the subsequent processing on the face is terminated (i.e., no subsequent operation is performed), and if the image includes a plurality of faces, the same feature extraction and matching processing is subsequently performed on each face respectively.
In the feature extraction step, a plurality of appearance features (or called depth features) of the human face are extracted through a neural network algorithm, and preferably, the appearance features are further segmented, encoded and classified through a semantic segmentation algorithm. Preferably, the appearance features include five sense organs (eyes, ears, nose, eyebrows, mouth), hairstyle, hair color, facial form, expression, and skin tone.
It can be understood that the above steps encode the features of five sense organs, hair style, hair color, face shape, expression and skin color corresponding to the facial appearance features, and are used as personalized building block image matching. For example, for hair style characteristics, segmenting an input face image into hair style parts by a semantic segmentation algorithm, and further encoding the characteristics of the hair style parts to measure hair styles of different types; for the hair color features, the colors of the hair parts are directly classified, the hair colors of the user are identified, and the features are used for the subsequent feature matching step.
In the step of feature matching, the closest image is extracted from a preset building block image library as a final matching result. The method comprises the steps of calculating the feature values of five sense organs, hair style, hair color, face shape, expression and skin color in the building block image library, performing feature similarity measurement calculation on the feature values and the appearance features of the input human face, and matching the building block image (preferably cartoon building block image) with the highest similarity with the input human face. Preferably, the feature similarity metric calculation may be a cosine similarity calculation, a pearson correlation coefficient calculation, or the like.
The invention also relates to a human image building block intelligent matching system based on the characteristic similarity measurement, which corresponds to the human image building block intelligent matching method based on the characteristic similarity measurement and can be understood as a system for realizing the method, and the system has a preferred structure as shown in fig. 3 and comprises a human face detection and recognition module, a characteristic extraction module and a characteristic matching module which are connected in sequence. The face detection and identification module is used for acquiring a target image and identifying the position of a face in the target image through a deep learning algorithm; the characteristic extraction module extracts a plurality of appearance characteristics of the face through a neural network algorithm, preferably, the appearance characteristics comprise characteristics of five sense organs, a hairstyle, a hair color, a face shape, an expression, a skin color and the like; for hair style characteristics, segmenting an input face image into hair style parts by a semantic segmentation algorithm, and further coding the characteristics of the hair style parts to measure hair styles of different types; for the hair color characteristics, directly classifying the colors of the hair part, identifying the hair colors, and using the characteristics in a next characteristic matching module; and the feature matching module is used for performing feature similarity measurement calculation on each extracted appearance feature of the human face and a corresponding building block image in a preset building block image library respectively, and automatically matching a building block image with the highest similarity with the input human face, wherein preferably, the similarity calculation comprises cosine similarity calculation and Pearson correlation coefficient calculation. Further, the feature matching module calculates corresponding feature values of five sense organs, hair style, hair color, face shape, expression and skin color for each image in the building block image library, performs feature similarity measurement calculation with the appearance features of the input face, and finally automatically matches a cartoon building block image with the highest similarity with the input face.
The system of the invention respectively completes the detection and cutting of human faces in input pictures, the extraction of human face characteristics and the matching with a building block image library through a human face detection and identification module, a characteristic extraction module and a characteristic matching module which are mutually connected and work cooperatively. The key of the scheme is a feature extraction module and a feature matching module, various features of a human face can be automatically extracted through a neural network model, feature similarity calculation can be carried out on the features and images in a building block image library, a cartoon building block image which is a cartoon portrait building block and is very similar to the human face can be automatically matched from the input human face, the personalization degree is high, matching from a human face picture to the portrait building block can be realized within second-level time, and the method is extremely high in efficiency.
It should be noted that the above-mentioned embodiments enable a person skilled in the art to more fully understand the invention, without restricting it in any way. Therefore, although the present invention has been described in detail with reference to the drawings and examples, it will be understood by those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention.
Claims (10)
1. An intelligent portrait building block matching method based on feature similarity measurement is characterized by comprising the following steps:
acquiring a target image, and identifying the position of a face in the target image through a deep learning algorithm;
a characteristic extraction step: extracting a plurality of appearance features of the human face through a neural network algorithm;
and (3) feature matching: and respectively carrying out feature similarity measurement calculation on each extracted appearance feature of the human face and the corresponding building block image in a preset building block image library, and automatically matching the building block image with the highest similarity with the input human face.
2. The intelligent human image building block matching method according to claim 1, wherein in the human face detection and recognition step, when the human face in the target image cannot be recognized through a deep learning algorithm, the operation is stopped; and if the target image contains a plurality of faces, respectively carrying out subsequent feature extraction and matching operation on each face.
3. The intelligent human image building block matching method as claimed in claim 2, wherein in the human face detection and recognition step, after the human face position is recognized, the gender of the human face is further judged through a deep learning algorithm, and for the human faces with different genders, the feature matching step is emphasized so as to match the human faces with the building block images conforming to the corresponding gender features.
4. The intelligent human image building block matching method as claimed in one of claims 1 to 3, wherein in the feature extraction step, after the appearance features are extracted, the appearance features are further segmented, encoded and classified through a semantic segmentation algorithm.
5. The intelligent human image building block matching method as claimed in claim 4, wherein in the feature extraction step, the appearance features include five sense organs, hair style, hair color, face shape, expression and skin color; for hair style characteristics, segmenting an input face image into hair style parts by a semantic segmentation algorithm, and further coding the characteristics of the hair style parts to measure hair styles of different types; for the hair color feature, the colors of the hair part are directly classified, and the hair color is identified.
6. The intelligent human image building block matching method according to claim 5, wherein in the feature matching step, corresponding feature values of five sense organs, hair style, hair color, face shape, expression and skin color are calculated for each image in the building block image library in the same way, feature similarity measurement calculation is performed with the appearance features of the input human face, and finally a cartoon building block image with the highest similarity with the input human face is automatically matched.
7. The intelligent portrait building block matching method according to claim 6, wherein in the feature matching step, the feature similarity metric calculation includes a cosine similarity calculation and a Pearson correlation coefficient calculation.
8. An intelligent portrait building block matching system based on feature similarity measurement is characterized by comprising a face detection and recognition module, a feature extraction module and a feature matching module which are sequentially connected;
the face detection and recognition module acquires a target image and recognizes the position of a face in the target image through a deep learning algorithm;
the feature extraction module extracts a plurality of appearance features of the human face through a neural network algorithm;
and the feature matching module is used for respectively carrying out feature similarity measurement calculation on each extracted appearance feature of the human face and the corresponding building block image in the preset building block image library, and automatically matching the building block image with the highest similarity with the input human face.
9. The intelligent human image building block matching system as claimed in claim 8, wherein the appearance features extracted by the feature extraction module include five sense organs, hair style, hair color, face shape, expression and skin color; for hair style characteristics, segmenting an input face image into hair style parts by a semantic segmentation algorithm, and further coding the characteristics of the hair style parts to measure hair styles of different types; for the hair color feature, the colors of the hair part are directly classified, and the hair color is identified.
10. The intelligent human image building block matching system according to claim 8, wherein the feature matching module calculates feature values of five sense organs, hair style, hair color, face shape, expression and skin color for each image in the building block image library, performs feature similarity measurement calculation with the appearance features of the input human face, the feature similarity measurement calculation includes cosine similarity calculation and pearson correlation coefficient calculation, and finally automatically matches the cartoon building block image with the highest face similarity with the input human face.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110816120.9A CN113269170A (en) | 2021-07-20 | 2021-07-20 | Intelligent portrait building block matching method and system based on feature similarity measurement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110816120.9A CN113269170A (en) | 2021-07-20 | 2021-07-20 | Intelligent portrait building block matching method and system based on feature similarity measurement |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113269170A true CN113269170A (en) | 2021-08-17 |
Family
ID=77236825
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110816120.9A Pending CN113269170A (en) | 2021-07-20 | 2021-07-20 | Intelligent portrait building block matching method and system based on feature similarity measurement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113269170A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116421960A (en) * | 2023-06-12 | 2023-07-14 | 山东中清智能科技股份有限公司 | Grabbing control method and device based on image processing |
| WO2023226858A1 (en) * | 2022-05-27 | 2023-11-30 | 北京字跳网络技术有限公司 | Method and apparatus for rendering virtual profile photo, and electronic device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105374055A (en) * | 2014-08-20 | 2016-03-02 | 腾讯科技（深圳）有限公司 | Image processing method and device |
| CN109308682A (en) * | 2018-10-18 | 2019-02-05 | 杭州数为科技有限公司 | A kind of recognition of face and cartoon restoring method, device, medium and system |
| CN110688962A (en) * | 2019-09-29 | 2020-01-14 | 武汉秀宝软件有限公司 | Face image processing method, user equipment, storage medium and device |
| US20200034604A1 (en) * | 2018-07-24 | 2020-01-30 | Sultan A. Alrasheed | Facial modelling and matching systems and methods |
| CN111260763A (en) * | 2020-01-21 | 2020-06-09 | 厦门美图之家科技有限公司 | Cartoon image generation method, device, equipment and storage medium based on portrait |
| CN111354079A (en) * | 2020-03-11 | 2020-06-30 | 腾讯科技（深圳）有限公司 | Three-dimensional face reconstruction network training and virtual face image generation method and device |
-
2021
- 2021-07-20 CN CN202110816120.9A patent/CN113269170A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105374055A (en) * | 2014-08-20 | 2016-03-02 | 腾讯科技（深圳）有限公司 | Image processing method and device |
| US20200034604A1 (en) * | 2018-07-24 | 2020-01-30 | Sultan A. Alrasheed | Facial modelling and matching systems and methods |
| CN109308682A (en) * | 2018-10-18 | 2019-02-05 | 杭州数为科技有限公司 | A kind of recognition of face and cartoon restoring method, device, medium and system |
| CN110688962A (en) * | 2019-09-29 | 2020-01-14 | 武汉秀宝软件有限公司 | Face image processing method, user equipment, storage medium and device |
| CN111260763A (en) * | 2020-01-21 | 2020-06-09 | 厦门美图之家科技有限公司 | Cartoon image generation method, device, equipment and storage medium based on portrait |
| CN111354079A (en) * | 2020-03-11 | 2020-06-30 | 腾讯科技（深圳）有限公司 | Three-dimensional face reconstruction network training and virtual face image generation method and device |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023226858A1 (en) * | 2022-05-27 | 2023-11-30 | 北京字跳网络技术有限公司 | Method and apparatus for rendering virtual profile photo, and electronic device and storage medium |
| CN116421960A (en) * | 2023-06-12 | 2023-07-14 | 山东中清智能科技股份有限公司 | Grabbing control method and device based on image processing |
| CN116421960B (en) * | 2023-06-12 | 2023-08-25 | 山东中清智能科技股份有限公司 | Grabbing control method and device based on image processing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2014368997B2 (en) | System and method for identifying faces in unconstrained media | |
| US9129191B2 (en) | Semantic object selection | |
| CN111160264B (en) | Cartoon character identity recognition method based on generation countermeasure network | |
| US9129192B2 (en) | Semantic object proposal generation and validation | |
| CN113269170A (en) | Intelligent portrait building block matching method and system based on feature similarity measurement | |
| CN110991389A (en) | Matching method for judging appearance of target pedestrian in non-overlapping camera view angle | |
| CN113705290A (en) | Image processing method, image processing device, computer equipment and storage medium | |
| KR100723417B1 (en) | Apparatuses and methods for recognizing face, and apparatus and method for extracting face from multiple face images | |
| CN109389076B (en) | Image segmentation method and device | |
| CN111126307A (en) | Small sample face recognition method of joint sparse representation neural network | |
| CN112150692A (en) | Access control method and system based on artificial intelligence | |
| CN113920472A (en) | Unsupervised target re-identification method and system based on attention mechanism | |
| CN112613416A (en) | Facial expression recognition method and related device | |
| CN109947960B (en) | Face multi-attribute joint estimation model construction method based on depth convolution | |
| CN110610138A (en) | Facial emotion analysis method based on convolutional neural network | |
| CN115862120B (en) | Face action unit identification method and equipment capable of decoupling separable variation from encoder | |
| Gupta et al. | Deep learning based identity verification in renaissance portraits | |
| CN111191549A (en) | Two-stage face anti-counterfeiting detection method | |
| Karungaru et al. | Face recognition in colour images using neural networks and genetic algorithms | |
| CN112528905B (en) | Image processing method, device and computer storage medium | |
| CN109376746A (en) | A kind of image identification method and system | |
| Meenpal et al. | Face recognition system based on principal components analysis and distance measures | |
| CN115761048A (en) | Face age editing method based on video time sequence | |
| CN112380966B (en) | Monocular iris matching method based on feature point re-projection | |
| CN111695507B (en) | Static gesture recognition method based on improved VGGNet network and PCA |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210817 |
|
| RJ01 | Rejection of invention patent application after publication |