CN109993100B - Method for realizing facial expression recognition based on deep feature clustering - Google Patents
Method for realizing facial expression recognition based on deep feature clustering Download PDFInfo
- Publication number
- CN109993100B CN109993100B CN201910240401.7A CN201910240401A CN109993100B CN 109993100 B CN109993100 B CN 109993100B CN 201910240401 A CN201910240401 A CN 201910240401A CN 109993100 B CN109993100 B CN 109993100B
- Authority
- CN
- China
- Prior art keywords
- facial expression
- network
- clustering
- loss function
- pictures
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000008921 facial expression Effects 0.000 title claims abstract description 101
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 28
- 238000001514 detection method Methods 0.000 claims abstract description 6
- 238000005259 measurement Methods 0.000 claims abstract description 4
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 230000014509 gene expression Effects 0.000 claims description 22
- 238000012549 training Methods 0.000 claims description 19
- 230000001815 facial effect Effects 0.000 claims description 6
- 238000003064 k means clustering Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 24
- 238000013135 deep learning Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000009193 crawling Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- 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
- G06N3/045—Combinations of networks
-
- 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]
- G06V10/464—Salient features, e.g. scale invariant feature transforms [SIFT] using a plurality of salient features, e.g. bag-of-words [BoW] representations
-
- 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
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships
-
- 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
-
- 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
- G06V40/174—Facial expression recognition
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Human Computer Interaction (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Geometry (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a method for realizing facial expression recognition based on deep feature clustering, which comprises the following steps: s1: collecting various facial expression pictures, and classifying the facial expression pictures one by one according to the facial expressions; s2: preprocessing the picture, removing the fuzzy picture, obtaining key points of the face by using a convolutional neural network-based cascading multitask face detection algorithm, and uniformly cutting the face picture according to the key points; s3: constructing a facial expression recognition network based on a convolutional neural network, and respectively inputting the preprocessed facial expression pictures into the network to calculate a loss function and train the loss function; s4: and acquiring the trained facial expression recognition network, and applying the trained facial expression recognition network to actual measurement. The method solves the problems of low accuracy rate, overfitting and the like of facial expression recognition.
Description
Technical Field
The invention relates to a method for realizing facial expression recognition based on deep feature clustering, which can be used in the technical field of computer vision picture processing.
Background
In recent years, with the rapid development of artificial intelligence, deep learning has become an area of intense research. Deep learning is excellent in solving many problems such as image object recognition, voice recognition, and natural language processing. Among various types of neural networks, convolutional neural networks are most intensively studied. Early on, due to the lack of training data and computational power, it was difficult to train a high performance convolutional neural network without generating overfitting. The emergence of large-scale labeled data such as ImageNet and the rapid improvement of GPU computational performance have led to rapid well-blowout for the study of convolutional neural networks.
With the continuous development of the convolutional neural network, the fitting analysis capability of the model to the real data is stronger and stronger, and meanwhile, in order to take speed and precision into consideration, researchers provide a plurality of lightweight convolutional neural networks. The lightweight convolutional neural network can achieve high reasoning speed, meanwhile achieve high accuracy rate, and fully utilize parameters of the network. The Mobilenet-V2 network is a lightweight convolutional neural network developed by Google, and is characterized by having fewer parameters and being capable of realizing real-time operation on a mobile phone.
The facial expression recognition belongs to fine-grained feature recognition, and the Mobilenet-V2 is directly applied to the facial expression recognition, so that the phenomenon of low recognition accuracy rate or overfitting is easily caused. For the fine-grained facial expression characteristics, how to make the network easily realize the accurate division of the expression is a technical problem which is urgently solved.
Disclosure of Invention
The invention aims to solve the problems in the prior art and provides a method for realizing facial expression recognition based on deep feature clustering.
The purpose of the invention is realized by the following technical scheme: the method for realizing facial expression recognition based on deep feature clustering comprises the following steps:
s1: collecting various facial expression pictures, and classifying the facial expression pictures one by one according to facial expressions to obtain a classified facial expression data set;
s2: preprocessing the classified facial expression data set pictures obtained in the step S1, removing fuzzy pictures, obtaining facial key points by using a convolutional neural network-based cascading multitask face detection algorithm, and uniformly cutting the facial pictures according to the key points to obtain a preprocessed facial expression data set;
s3: constructing a facial expression recognition network based on a convolutional neural network, and respectively inputting the preprocessed facial expression data set pictures obtained in the step S2 into the network to calculate a loss function and train the loss function to obtain a trained facial expression recognition network;
s4: and (4) the trained facial expression recognition network obtained in the step S3 is applied to actual measurement.
Preferably, in the step S1, the collected facial expression pictures need to be of balanced category, each type of facial expression picture needs to be at least two thousand facial expression pictures, and the face needs to be clear and the posture needs to be correct.
Preferably, in the step S2, the image is preprocessed to remove the blurred image, then the face key points are obtained by using a convolutional neural network-based cascaded multi-task face detection algorithm, the face images are uniformly cut according to the key points, and then the face images are respectively saved according to the face expressions, if there are few face expression images, data enhancement is performed on the face images.
Preferably, in the step S3, the convolutional neural network has a structure of mobilene-V2, and the input layer is the clipped face picture and outputs probability values of various facial expressions.
Preferably, in the step S3, deep feature clustering loss is added to the loss function of the convolutional neural network, so that the deep features obtained by various classes of facial expression pictures through the convolutional neural network have larger differences.
Preferably, the training of the facial expression recognition algorithm based on deep feature clustering in the step S3 includes the steps of:
s31: inputting the facial expression data preprocessed in the step S2 into a pre-trained Mobilene-V2 network in sequence according to the expression categories, extracting high latitude features of a penultimate layer 1280 x 1 in the network in sequence, clustering the high latitude features of each category of expressions by adopting a K-means clustering algorithm to obtain K clustering centers of each facial expression, and iteratively updating the clustering centers once in each cycle;
s32: comparing the K clustering centers of each facial expression in the step S31 with the high latitude characteristics of the same layer of each training sample to obtain a clustering loss function;
s33: training a convolutional neural network model to minimize a loss function of the network
Preferably, the loss function in the step S3 is designed to be
Wherein,
L k-means (f,a,c)=||max(f,c a )-min(f,c -a )||
wherein L in the formula is a total loss function,for categorizing the cross-entropy loss function, L k-means (f, a, c) is a clustering loss function, x is an input facial expression training image, a is a facial expression label corresponding to the input image x,obtaining a predicted label of an input image x through a Mobilene-V2 network, f obtaining high latitude characteristics of a penultimate layer 1280 x 1 of the input image x through a Mobilene-V2 network, c obtaining K clustering centers of N expressions after all the high latitude characteristics of a training picture are clustered through a pre-trained Mobilene-V2 network, wherein the K clustering centers are N x K clustering centers in total, and c is a K cluster centers with expression a, c -a There are (N-1) × K cluster centers for all expressions except expression a.
Compared with the prior art, the invention adopting the technical scheme has the following technical effects: the invention adopts a method of enlarging the distance between the deep image features, thereby enabling the network to easily realize accurate division of the expression. The facial expression recognition algorithm based on deep feature clustering can enlarge the distance between deep features of facial expression pictures in a Mobilenet-V2 network, so that fine-grained facial expression classification is more accurate. The method solves the problems of low accuracy rate, overfitting and the like of facial expression recognition.
Drawings
Fig. 1 is a structural diagram of mobilene-V2 in the deep feature cluster-based facial expression recognition algorithm of the present invention.
FIG. 2 is a structural diagram of a residual network block in the deep feature clustering-based facial expression recognition algorithm of the present invention.
Detailed Description
Objects, advantages and features of the present invention will be illustrated and explained by the following non-limiting description of preferred embodiments. The embodiments are merely exemplary for applying the technical solutions of the present invention, and any technical solution formed by replacing or converting the equivalent thereof falls within the scope of the present invention claimed.
The invention discloses a method for realizing facial expression recognition based on deep feature clustering, which comprises the following steps:
s1: various facial expression pictures are collected and classified one by one according to facial expressions.
The method comprises the following specific steps: and finding a picture website, finding a facial expression picture and ensuring that the picture is relatively clear. And (3) crawling various facial expression pictures from the website by using a crawler technology, and ensuring that each type of facial expression picture is more than twenty thousand.
S2: preprocessing the picture, removing the fuzzy picture, obtaining five key points of the face by using a convolutional neural network-based cascade multitask face detection algorithm (MTCNN), and uniformly cutting the face picture according to the key points.
And screening the pictures one by one, and removing the pictures with blurs and inconsistent picture contents. And uniformly cutting the screened pictures into 128-by-128 sizes, and respectively storing according to various expressions of the face images.
S3: and constructing a facial expression recognition network based on a convolutional neural network, and respectively inputting the preprocessed facial expression pictures into the network to calculate a loss function and train the loss function.
The network structure of the Mobilenet-V2 is shown in fig. 1. Mobilenet-V2 is composed of four parts: convolutional layer, global average pooling layer, residual network block. The convolutional layer extracts feature information of the picture through convolution operation, and the extracted information is more and more abstract along with the multi-layer superposition of the convolution operation. Residual network blocks in the network structure are shown in fig. 2, and the residual network blocks are used for transferring the bottom layer features into the high layer and inhibiting the situation of gradient disappearance. The input to the mobilent-V2 is a facial expression picture, and the predicted facial expression label is output.
The loss function is composed of a classification cross entropy loss function and a clustering loss function. The classification cross entropy loss function is used for improving the classification accuracy of the network, and the clustering loss function is used for enlarging the high latitude characteristic difference generated by different types of facial expression images through the network.
And adding deep feature clustering loss into the loss function of the convolutional neural network in the step S3, so that the deep features obtained by the facial expression pictures of various categories through the convolutional neural network have larger difference, which is beneficial to distinguishing fine-grained facial features and fine-grained facial features.
The training process specifically comprises the following steps:
s31: and (4) sequentially inputting the preprocessed facial expression data in the step (S2) into a pre-trained Mobilene-V2 network according to the expression classes, sequentially extracting high latitude features of the 1 x 1280 on the penultimate layer in the network, and clustering the high latitude features of the N types of expressions by adopting a K-means clustering algorithm to obtain K clustering centers (clusters) of each facial expression, wherein the N clusters are N x K clusters.
S32: and comparing the N x K clustering centers in the step of S31 with the high latitude characteristics of the same layer of each training sample to obtain a clustering loss function. And calculating 1 x 1280 latitude characteristics of the input facial expression picture during training, finding out the distance between the same type expression cluster farthest from the characteristics and the nearest non-same type expression cluster, and then respectively calculating the distance between the characteristics and the two clusters. The difference of the two distances, i.e. the cluster loss function, is maximized. And after all the training pictures are trained for one round, the network model is stored, N x K clusters are recalculated, and iterative training is performed again.
S33: the convolutional neural network model is trained to minimize a loss function of the network.
The loss function is:
wherein,
L k-means (f,a,c)=||max(f,c a )-min(f,c -a )||
in the formula L is the overall loss function,for categorizing the cross-entropy loss function, L k-means (f, a, c) is a clustering loss function, x is an input facial expression training image, a is a facial expression label corresponding to the input image x,obtaining a predicted label of an input image x through a Mobilene-V2 network, f obtaining high latitude characteristics of a penultimate layer 1 x 1280 of the input image x through a Mobilene-V2 network, c obtaining K clustering centers (N x K clustering centers in total) of N expressions after all the high latitude characteristics of training pictures are clustered through a pre-trained Mobilene-V2 network, and c obtaining K clustering centers of the N expressions (N x K clustering centers in total) through pre-training images a K cluster centers expressed as a, c -a K cluster centers for all expressions except expression a (total of (N-1) × K cluster centers).
S4: and acquiring a trained facial expression recognition network, and applying the trained facial expression recognition network to actual measurement.
In conclusion, the invention can obtain a face recognition network model with both precision and speed, and the generalization capability of the network is strong. The invention obtains the trained facial expression recognition network by inputting the facial expression pictures into the Mobilene-V2 network and training the model through the facial expression recognition algorithm based on deep feature clustering. At the moment, the network can better identify fine-grained facial expressions. The facial expression recognition algorithm based on deep feature clustering is applied to facial expression recognition, so that the difference between facial expression classes is enlarged, and the problem that fine-grained images are difficult to recognize is optimized.
The invention has various embodiments, and all technical solutions formed by adopting equivalent transformation or equivalent transformation are within the protection scope of the invention.
Claims (6)
1. The method for realizing facial expression recognition based on deep feature clustering is characterized by comprising the following steps: the method comprises the following steps:
s1: collecting various facial expression pictures, and classifying the facial expression pictures one by one according to facial expressions to obtain a classified facial expression data set;
s2: preprocessing the classified facial expression data set pictures obtained in the step S1, removing fuzzy pictures, obtaining facial key points by using a convolutional neural network-based cascading multitask face detection algorithm, and uniformly cutting the facial pictures according to the key points to obtain a preprocessed facial expression data set;
s3: constructing a facial expression recognition network based on a convolutional neural network, and respectively inputting the preprocessed facial expression data set pictures obtained in the step S2 into the network to calculate a loss function and train the loss function to obtain a trained facial expression recognition network;
the loss function is designed as
Wherein,
L k-means (f,a,c)=||max(f,c a )-min(f,c -a )||
wherein L in the formula is a total loss function,for categorizing the cross-entropy loss function, L k-means (f, a, c) is clusteringA loss function, x is an input facial expression training image, a is a facial expression label corresponding to the input image x,obtaining a predicted label of an input image x through a Mobilene-V2 network, f obtaining a high-dimensional feature of a penultimate layer 1280 x 1 of the input image x through a Mobilene-V2 network, c obtaining K clustering centers of N expressions after all the high-dimensional features of a training picture are clustered through a pre-trained Mobilene-V2 network, wherein the K clustering centers are N x K clustering centers in total, and c is a K cluster centers expressed as a, c -a The total number of the cluster centers of all expressions except the expression a is (N-1) × K;
s4: and identifying the trained facial expression obtained in the step S3, and applying the identified network to actual measurement.
2. The method for realizing facial expression recognition based on deep feature clustering according to claim 1, wherein the method comprises the following steps: in the step S1, the collected facial expression pictures need to be of a balanced category, each facial expression picture needs to be at least two thousand facial expression pictures, and the face needs to be clear and the posture needs to be correct.
3. The method for realizing facial expression recognition based on deep feature clustering according to claim 1, wherein the method comprises the following steps: in the step S2, the image is preprocessed to remove the blurred image, then the cascaded multi-task face detection algorithm based on the convolutional neural network is used to obtain face key points, the face image is uniformly cut according to the key points, and then the face image is respectively stored according to the face expression, if there is less face expression images, the data enhancement is performed on the face image.
4. The method for realizing facial expression recognition based on deep feature clustering according to claim 1, wherein the method comprises the following steps: in the step S3, the convolutional neural network structure is mobilene-V2, and the input layer is the clipped human face picture and outputs probability values of various human face expressions.
5. The method for realizing facial expression recognition based on deep feature clustering according to claim 1, wherein the method comprises the following steps: in the step S3, deep feature clustering loss is added to the loss function of the convolutional neural network, so that the deep feature differences obtained by the various types of facial expression pictures through the convolutional neural network are larger.
6. The method for realizing facial expression recognition based on deep feature clustering according to claim 1, wherein the method comprises the following steps: training a facial expression recognition algorithm based on deep feature clustering in the step S3, including the steps of:
s31: inputting the facial expression data preprocessed in the step S2 into a pre-trained Mobilene-V2 network in sequence according to expression categories, extracting high-dimensional features of the last but one layer 1280 x 1 in the network in sequence, clustering the high-dimensional features of each type of expression by adopting a K-means clustering algorithm to obtain K clustering centers of each facial expression, and iteratively updating the clustering centers once in each cycle;
s32: comparing the K clustering centers of each facial expression in the step S31 with the high-dimensional features of the same layer of each training sample to obtain a clustering loss function;
s33: the convolutional neural network model is trained to minimize the loss function of the network.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910240401.7A CN109993100B (en) | 2019-03-27 | 2019-03-27 | Method for realizing facial expression recognition based on deep feature clustering |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910240401.7A CN109993100B (en) | 2019-03-27 | 2019-03-27 | Method for realizing facial expression recognition based on deep feature clustering |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109993100A CN109993100A (en) | 2019-07-09 |
CN109993100B true CN109993100B (en) | 2022-09-20 |
Family
ID=67131863
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910240401.7A Active CN109993100B (en) | 2019-03-27 | 2019-03-27 | Method for realizing facial expression recognition based on deep feature clustering |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109993100B (en) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110569878B (en) * | 2019-08-08 | 2022-06-24 | 上海汇付支付有限公司 | Photograph background similarity clustering method based on convolutional neural network and computer |
CN110781784A (en) * | 2019-10-18 | 2020-02-11 | 高新兴科技集团股份有限公司 | Face recognition method, device and equipment based on double-path attention mechanism |
CN111126244A (en) * | 2019-12-20 | 2020-05-08 | 南京邮电大学 | Security authentication system and method based on facial expressions |
CN111401193B (en) * | 2020-03-10 | 2023-11-28 | 海尔优家智能科技(北京)有限公司 | Method and device for acquiring expression recognition model, and expression recognition method and device |
CN111414862B (en) * | 2020-03-22 | 2023-03-24 | 西安电子科技大学 | Expression recognition method based on neural network fusion key point angle change |
CN111507224B (en) * | 2020-04-09 | 2022-08-30 | 河海大学常州校区 | CNN facial expression recognition significance analysis method based on network pruning |
CN112232116A (en) * | 2020-09-08 | 2021-01-15 | 深圳微步信息股份有限公司 | Facial expression recognition method and device and storage medium |
CN113033374A (en) * | 2021-03-22 | 2021-06-25 | 开放智能机器(上海)有限公司 | Artificial intelligence dangerous behavior identification method and device, electronic equipment and storage medium |
CN113076930B (en) * | 2021-04-27 | 2022-11-08 | 东南大学 | Face recognition and expression analysis method based on shared backbone network |
CN117542106B (en) * | 2024-01-10 | 2024-04-05 | 成都同步新创科技股份有限公司 | Static face detection and data elimination method, device and storage medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108304826A (en) * | 2018-03-01 | 2018-07-20 | 河海大学 | Facial expression recognizing method based on convolutional neural networks |
CN108764207A (en) * | 2018-06-07 | 2018-11-06 | 厦门大学 | A kind of facial expression recognizing method based on multitask convolutional neural networks |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190042952A1 (en) * | 2017-08-03 | 2019-02-07 | Beijing University Of Technology | Multi-task Semi-Supervised Online Sequential Extreme Learning Method for Emotion Judgment of User |
-
2019
- 2019-03-27 CN CN201910240401.7A patent/CN109993100B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108304826A (en) * | 2018-03-01 | 2018-07-20 | 河海大学 | Facial expression recognizing method based on convolutional neural networks |
CN108764207A (en) * | 2018-06-07 | 2018-11-06 | 厦门大学 | A kind of facial expression recognizing method based on multitask convolutional neural networks |
Also Published As
Publication number | Publication date |
---|---|
CN109993100A (en) | 2019-07-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109993100B (en) | Method for realizing facial expression recognition based on deep feature clustering | |
CN112308158B (en) | Multi-source field self-adaptive model and method based on partial feature alignment | |
CN107480261B (en) | Fine-grained face image fast retrieval method based on deep learning | |
CN108133188B (en) | Behavior identification method based on motion history image and convolutional neural network | |
CN112418117B (en) | Small target detection method based on unmanned aerial vehicle image | |
Wang et al. | Deep learning algorithms with applications to video analytics for a smart city: A survey | |
Asadi et al. | A convolution recurrent autoencoder for spatio-temporal missing data imputation | |
CN110321805B (en) | Dynamic expression recognition method based on time sequence relation reasoning | |
Xia et al. | Transferring ensemble representations using deep convolutional neural networks for small-scale image classification | |
CN112329536A (en) | Single-sample face recognition method based on alternative pair anti-migration learning | |
Menaga et al. | Deep learning: a recent computing platform for multimedia information retrieval | |
Zhang et al. | Classification of canker on small datasets using improved deep convolutional generative adversarial networks | |
CN116152554A (en) | Knowledge-guided small sample image recognition system | |
CN115458174A (en) | Method for constructing intelligent diagnosis model of diabetic retinopathy | |
Dhawan et al. | Deep Learning Based Sugarcane Downy Mildew Disease Detection Using CNN-LSTM Ensemble Model for Severity Level Classification | |
CN114048843A (en) | Small sample learning network based on selective feature migration | |
Li | Parallel two-class 3D-CNN classifiers for video classification | |
CN112434145A (en) | Picture-viewing poetry method based on image recognition and natural language processing | |
CN116051924B (en) | Divide-and-conquer defense method for image countermeasure sample | |
Dhanalakshmi et al. | Tomato leaf disease identification by modified inception based sequential convolution neural networks | |
CN116797821A (en) | Generalized zero sample image classification method based on fusion visual information | |
CN115374943A (en) | Data cognition calculation method and system based on domain confrontation migration network | |
CN115019342A (en) | Endangered animal target detection method based on class relation reasoning | |
CN115100694A (en) | Fingerprint quick retrieval method based on self-supervision neural network | |
Guzzi et al. | Distillation of a CNN for a high accuracy mobile face recognition system |
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 |