CN107896362A - A kind of WIFI location fingerprints localization method and system based on deep learning - Google Patents

A kind of WIFI location fingerprints localization method and system based on deep learning Download PDF

Info

Publication number
CN107896362A
CN107896362A CN201711015262.5A CN201711015262A CN107896362A CN 107896362 A CN107896362 A CN 107896362A CN 201711015262 A CN201711015262 A CN 201711015262A CN 107896362 A CN107896362 A CN 107896362A
Authority
CN
China
Prior art keywords
classification
rssi information
wifi
deep learning
training
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
Application number
CN201711015262.5A
Other languages
Chinese (zh)
Inventor
高建彬
张丽园
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN201711015262.5A priority Critical patent/CN107896362A/en
Publication of CN107896362A publication Critical patent/CN107896362A/en
Pending legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • H04W64/006—Locating users or terminals or network equipment for network management purposes, e.g. mobility management with additional information processing, e.g. for direction or speed determination
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0252—Radio frequency fingerprinting
    • 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

Abstract

The invention discloses a kind of WIFI location fingerprints localization method and system based on deep learning, it is related to the WIFI localization methods field based on deep learning；It comprises the following steps：1) obtain RSSI information using WIFI adapters scanning WIFI signal and be inputted progress unsupervised learning in parameter training and classification and orientation unit and obtain reconstructing RSSI information；2) judge whether reconstruct RSSI information is consistent with RSSI information, if unanimously, then terminating parameter training, skips to step 3；3) by the study that exercised supervision in the training of reconstruct RSSI information input parameter and classification and orientation unit, complete classification and orientation and obtain the affiliated building of each RSSI information sources and floor completion positioning；The present invention solve the problems, such as it is existing realize that WIFI location fingerprints positioning high accuracy needs later stage numerous and diverse time-consuming optimization and over-fitting occurs based on simple deep learning, reached the effect for the precision for improving the WIFI location fingerprints positioning based on deep learning.

Description

A kind of WIFI location fingerprints localization method and system based on deep learning
Technical field
The present invention relates to the WIFI localization methods field based on deep learning, especially a kind of WIFI based on deep learning Location fingerprint localization method and system.
Background technology
Location fingerprint, which refers to, connects the position in actual environment and certain " fingerprint ", one uniqueness of a position correspondence Fingerprint；Fingerprint can be one-dimensional or multidimensional, for example equipment to be positioned is receiving or sent information, then fingerprint can be This information or a feature of signal or multiple features, most commonly signal intensity；WIFI is nearly ubiquitous to be can use Property becomes a very attractive localization method, without extra hardware spending, the positioning side based on time and angle Method is not suitable for WIFI signal so that location fingerprint method, which turns into, positions main selection.
For deep learning, its thought is exactly to stack the input exported as next layer of this i.e. layer of multiple layers； In this way, it is possible to realize that carrying out hierarchical table to input information reaches；The artificial neural network of more hidden layers has excellent Feature learning ability, the feature for learning to obtain have to data it is more essential portray, so as to be advantageous to visualize or classify；Depth god Difficulty through network in training, can effectively it be overcome by " successively initializing ", successively initialization can be by unsupervised Study is realized；The WIFI localization method generally use deep learning networks of deep learning are currently based on, the later stage, which optimizes, to be reached It is accurately positioned, wherein deep learning network uses existing convolutional layer+pond layer+classification layer, and the solution of optimization used Filter, manual data analysis and time-consuming parameter adjustment, it is low to implement numerous and diverse, time-consuming and efficiency；On the other hand due to RSSI's Sampled data is the discrete values using the time as dimension, if using deep learning algorithm be easy to occur over-fitting situation cause it is fixed The low precision of position even positions failure.
The content of the invention
It is an object of the invention to：The invention provides a kind of WIFI location fingerprints localization method based on deep learning and System, solve and existing realize that WIFI location fingerprints positioning high accuracy needs the later stage numerous and diverse time-consuming excellent based on simple deep learning The problem of changing and over-fitting occurs, the effect for the precision for improving the WIFI location fingerprints positioning based on deep learning is reached Fruit.
The technical solution adopted by the present invention is as follows：
A kind of WIFI location fingerprint localization methods based on deep learning, comprise the following steps：
Step 1：RSSI information is obtained using WIFI adapters scanning WIFI signal and is inputted parameter training and classification Unsupervised learning is carried out in positioning unit to obtain reconstructing RSSI information；
Step 2：Judge whether reconstruct RSSI information is consistent with RSSI information, if unanimously, then terminating parameter training, skips to Step 3；If inconsistent, step 1 is skipped to；
Step 3：By the study that exercised supervision in the training of reconstruct RSSI information input parameter and classification and orientation unit, classification is completed Positioning obtains the affiliated building of each RSSI information sources and floor completes positioning.
Preferably, unsupervised learning includes the study that parameter training and initial data reduction represent in the step 1.
Preferably, the step 3 comprises the following steps：
Step 3.1：After carrying out Dropout operations between parameter training and two hidden layers of classification and orientation unit, by weight Structure RSSI information is randomly divided into training set, active set and test set；
Step 3.2：Active set and training set are input to exercise supervision in parameter training and classification and orientation unit and learnt Hidden layer is adjusted after to classification results and includes neuron number to optimal；
Step 3.3：By test set be input in classification and orientation unit the study that exercises supervision obtain adjusting after classification results it is hidden Layer is comprising neuron number to optimal；
Step 3.4：After the optimal neuron number of step 3.2 and step 3.3 selection sort result, normalization is completed Operation show that the affiliated building of each RSSI information sources and floor complete positioning.
A kind of WIFI location fingerprint alignment systems based on deep learning, including WIFI adapters and parameter training and classification Positioning unit；
WIFI adapters, the signal intensity i.e. RSSI information of each signal is obtained for scanning WIFI signal；
Parameter training and classification and orientation unit, unsupervised learning output reconstruct RSSI information is carried out for inputting RSSI information Afterwards, and input reconstruct RSSI information and carry out classification and orientation and draw positioning result.
Preferably, parameter training and classification and orientation the unit network includes encoder, decoder and grader；Encoder with Decoder is connected to parameter training, and encoder is connected to classification and orientation with grader.
Preferably, the encoder includes an input layer and three hidden layers；The decoder includes two hidden layers and one Individual output layer；Information is sequentially input and information is output into output layer after three hidden layers and two hidden layers by the input layer.
Preferably, the grader includes two final hidden layers, a Dropout layer and a final output layer；Two Transmitted between final hidden layer by a Dropout layer, final hidden layer output is connected with final output layer.
In summary, by adopting the above-described technical solution, the beneficial effects of the invention are as follows：
1. the present invention carries out classification and orientation using deep neural network system architecture to building and floor, pretreatment net is set up Network reconstructs RSSI information, prevents over-fitting situation, is simplified while realizing and improve positioning precision based on deep learning WIFI location fingerprints positioning flow, solve it is existing based on simple deep learning realize WIFI location fingerprints positioning high accuracy need The problem of wanting later stage numerous and diverse time-consuming optimization and over-fitting occur；
2. the present invention uses stack autocoder, simple in construction, simple operation, manual debugging greatly reducing Workload, functional space is efficiently reduced, can realize powerful and accurately classify automatically；
3. the deep-neural-network of the present invention includes encoder (convolution+pond) and decoder (up-sampling+convolution), coding Device (convolution+pond) and grader two parts, the RSSI signals of the reconstruct to being obtained after signal transacting will not produce over-fitting, disappear Except some side effects, positioning precision is improved.
Brief description of the drawings
Examples of the present invention will be described by way of reference to the accompanying drawings, wherein：
Fig. 1 is deep neural network structure chart of the present invention for parameter training part；
Fig. 2 is deep neural network structure chart of the present invention for classified part.
Embodiment
All features disclosed in this specification, or disclosed all methods or during the step of, except mutually exclusive Feature and/or step beyond, can combine in any way.
The present invention is elaborated with reference to Fig. 1-2.
Embodiment 1
A kind of WIFI location fingerprint localization methods based on deep learning, comprise the following steps：
Step 1：RSSI information is obtained using WIFI adapters scanning WIFI signal and is inputted parameter training and classification Unsupervised learning is carried out in positioning unit to obtain reconstructing RSSI information；
Unsupervised learning includes the study that parameter training and initial data reduction represent in the step 1；
Step 2：Judge whether reconstruct RSSI information is consistent with RSSI information, if unanimously, then terminating parameter training, skips to Step 3；If inconsistent, step 1 is skipped to；
Step 3：By the study that exercised supervision in the training of reconstruct RSSI information input parameter and classification and orientation unit, classification is completed Positioning obtains the affiliated building of each RSSI information sources and floor completes positioning.
The step 3 comprises the following steps：
Step 3.1：After carrying out Dropout operations between parameter training and two hidden layers of classification and orientation unit, by weight Structure RSSI information is randomly divided into training set, active set and test set；
Step 3.2：Active set and training set are input to exercise supervision in parameter training and classification and orientation unit and learnt Hidden layer is adjusted after to classification results and includes neuron number to optimal；
Step 3.3：By test set be input in classification and orientation unit the study that exercises supervision obtain adjusting after classification results it is hidden Layer is comprising neuron number to optimal；
Step 3.4：After the optimal neuron number of step 3.2 and step 3.3 selection sort result, normalization is completed Operation show that the affiliated building of each RSSI information sources and floor complete positioning.
A kind of WIFI location fingerprint alignment systems based on deep learning, including WIFI adapters, parameter training network and Classification and orientation network；
WIFI adapters, the signal intensity i.e. RSSI information of each signal is obtained for scanning WIFI signal；
Parameter training and classification and orientation unit, unsupervised learning output reconstruct RSSI information is carried out for inputting RSSI information Afterwards, and input reconstruct RSSI information and carry out classification and orientation and draw positioning result.
Parameter training and classification and orientation unit network include encoder, decoder and grader；Encoder is connected with decoder For parameter training, encoder is connected to classification and orientation with grader.
Encoder includes an input layer and three hidden layers；Decoder includes two hidden layers and an output layer；Input layer Information is sequentially input information is output to output layer after three hidden layers and two hidden layers.
Grader includes two final hidden layers, a Dropout layer and a final output layer；Between two final hidden layers Transmitted by a Dropout layer, final hidden layer output is connected with final output layer.
Embodiment 2
Step 1：WIFI scan datas are obtained from 25 different Android devices or WIFI adapters, in database Each scanning includes 529 attributes；520 different AP are wherein shared, preceding 520 attributes represent the reception signal of these networks Intensity is RSSI information, and change in signal strength receives bad situation from 104dBm to 0dBm；When AP is unavailable, default value For 100；Remaining 9 attributes include longitude and latitude determination, floor quantity, building ID, space ID, relative position, ID Information, mobile phone ID and timestamp；
Step 2：The 520 bars intensity i.e. RSSI information scanned is input to deep neural network i.e. DNN structure charts In, the parameter of encoder section is trained, is terminated when the reconstruct RSSI information of output is consistent with the information inputted；On the other hand, While training parameter, learn initial data simplifies expression, can be by input feature vector space dimension according to different occasion needs Degree is reduced to 256,128 and 64 from 520；
Step 3：After carrying out Dropout operations between parameter training and two hidden layers of classification and orientation unit, it will reconstruct RSSI information is randomly divided into training set, active set and test set；Active set and training set are input to depth as shown in Figure 2 first It is the study that exercised supervision in DNN structures to spend neutral net, according to last classification results, respectively by encoder and grader two Point characteristic dimension is combined between 256,128 and 64 to be adjusted to optimal, that is, is adjusted hidden layer and included neuron number to optimal；
Step 4：Test set is input in deep neural network as shown in Figure 2 i.e. DNN structures the study that exercises supervision, root According to last classification results, encoder and grader two parts characteristic dimension are combined between 256,128 and 64 respectively Adjust to optimal, that is, adjust hidden layer and include neuron number to optimal.
Step 5：Weighed according to the result that step 3 and step 4 are different on data set, final choice classification results are most Excellent neuron number, obtain the best DNN structures of optimal generalization and complete normalization operation output result, it is fixed to complete Position.The present invention carries out classification and orientation using deep neural network system architecture to building and floor, sets up pretreatment network i.e. weight Structure RSSI information, over-fitting situation is prevented, the WIFI positions based on deep learning are simplified while realizing and improve positioning precision The flow of fingerprint location, solve it is existing based on simple deep learning realize WIFI location fingerprints positioning high accuracy need the later stage numerous Miscellaneous time-consuming optimization and the problem of over-fitting occurs.

Claims (7)

  1. A kind of 1. WIFI location fingerprint localization methods based on deep learning, it is characterised in that：Comprise the following steps：
    Step 1：RSSI information is obtained using WIFI adapters scanning WIFI signal and is inputted parameter training and classification and orientation Unsupervised learning is carried out in unit to obtain reconstructing RSSI information；
    Step 2：Judge whether reconstruct RSSI information is consistent with RSSI information, if unanimously, then terminating parameter training, skips to step 3；If inconsistent, step 1 is skipped to；
    Step 3：By the study that exercised supervision in the training of reconstruct RSSI information input parameter and classification and orientation unit, classification and orientation is completed Obtain the affiliated building of each RSSI information sources and floor completes positioning.
  2. A kind of 2. localization method of WIFI location fingerprints based on deep learning according to claim 1, it is characterised in that： Unsupervised learning includes the study that parameter training and initial data reduction represent in the step 1.
  3. A kind of 3. WIFI location fingerprint localization methods based on deep learning according to claim 1, it is characterised in that：Institute Step 3 is stated to comprise the following steps：
    Step 3.1：After carrying out Dropout operations between parameter training and two hidden layers of classification and orientation unit, it will reconstruct RSSI information is randomly divided into training set, active set and test set；
    Step 3.2：Active set and training set are input in parameter training and classification and orientation unit into the study that exercises supervision to be divided Hidden layer is adjusted after class result and includes neuron number to optimal；
    Step 3.3：Test set is input in classification and orientation unit after the study that exercises supervision obtains classification results and adjusts hidden layer bag Containing neuron number to optimal；
    Step 3.4：After the optimal neuron number of step 3.2 and step 3.3 selection sort result, normalization operation is completed Show that the affiliated building of each RSSI information sources and floor complete positioning.
  4. A kind of 4. WIFI location fingerprint alignment systems based on deep learning, it is characterised in that：Including WIFI adapters and parameter Training and classification and orientation unit；
    WIFI adapters, the signal intensity i.e. RSSI information of each signal is obtained for scanning WIFI signal；
    Parameter training and classification and orientation unit, after inputting RSSI information progress unsupervised learning output reconstruct RSSI information, And input reconstruct RSSI information progress classification and orientations and draw positioning result.
  5. A kind of 5. WIFI location fingerprint alignment systems based on deep learning according to claim 4, it is characterised in that：Institute Stating parameter training and classification and orientation unit includes encoder, decoder and grader；Encoder is connected to parameter with decoder Training, encoder are connected to classification and orientation with grader.
  6. A kind of 6. WIFI location fingerprint alignment systems based on deep learning according to claim 5, it is characterised in that：Institute Stating encoder includes an input layer and three hidden layers；The decoder includes two hidden layers and an output layer；The input Information is sequentially input and information is output into output layer after three hidden layers and two hidden layers by layer.
  7. A kind of 7. alignment system of WIFI location fingerprints based on deep learning according to claim 6, it is characterised in that： The grader includes two final hidden layers, a Dropout layer and a final output layer；Pass through between two final hidden layers One Dropout layers transmission, final hidden layer output are connected with final output layer.
CN201711015262.5A 2017-10-25 2017-10-25 A kind of WIFI location fingerprints localization method and system based on deep learning Pending CN107896362A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711015262.5A CN107896362A (en) 2017-10-25 2017-10-25 A kind of WIFI location fingerprints localization method and system based on deep learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711015262.5A CN107896362A (en) 2017-10-25 2017-10-25 A kind of WIFI location fingerprints localization method and system based on deep learning

Publications (1)

Publication Number Publication Date
CN107896362A true CN107896362A (en) 2018-04-10

Family

ID=61802891

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711015262.5A Pending CN107896362A (en) 2017-10-25 2017-10-25 A kind of WIFI location fingerprints localization method and system based on deep learning

Country Status (1)

Country Link
CN (1) CN107896362A (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108966131A (en) * 2018-07-23 2018-12-07 广州都市圈网络科技有限公司 Fusion floor estimating method based on indoor positioning
CN109040957A (en) * 2018-08-14 2018-12-18 广东小天才科技有限公司 A kind of indoor orientation method and device based on WIFI
CN109242097A (en) * 2018-07-12 2019-01-18 上海交通大学 The visual representation learning system and method for unsupervised learning
CN109246608A (en) * 2018-11-16 2019-01-18 重庆小富农康农业科技服务有限公司 A kind of point-to-point localization method in interior based on WIFI location fingerprint big data analysis
CN109359725A (en) * 2018-10-24 2019-02-19 北京周同科技有限公司 Training method, device, equipment and the computer readable storage medium of convolutional neural networks model
CN109474887A (en) * 2018-12-05 2019-03-15 哈尔滨工业大学 Height floor differentiating method based on LTE signal
CN110231593A (en) * 2019-04-11 2019-09-13 深圳市城市交通规划设计研究中心有限公司 Indoor orientation method, device, computer readable storage medium and terminal device
CN111031477A (en) * 2019-12-25 2020-04-17 中山大学 Millimeter wave indoor positioning method based on deep learning
CN114514759A (en) * 2019-10-04 2022-05-17 三星电子株式会社 System and method for WiFi-based indoor positioning via unsupervised domain adaptation

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160084646A1 (en) * 2014-09-24 2016-03-24 Wistron Corp. Methods for indoor positioning and apparatuses using the same
CN106658422A (en) * 2017-01-09 2017-05-10 深圳市智开科技有限公司 Network side positioning method and network side positioning system for aiming at high-sparse WiFi data
CN107064913A (en) * 2017-03-10 2017-08-18 上海斐讯数据通信技术有限公司 A kind of wireless location method and system based on deep learning

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160084646A1 (en) * 2014-09-24 2016-03-24 Wistron Corp. Methods for indoor positioning and apparatuses using the same
CN106658422A (en) * 2017-01-09 2017-05-10 深圳市智开科技有限公司 Network side positioning method and network side positioning system for aiming at high-sparse WiFi data
CN107064913A (en) * 2017-03-10 2017-08-18 上海斐讯数据通信技术有限公司 A kind of wireless location method and system based on deep learning

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
刘震宇，李嘉俊，王昆: "一种基于深度自编码器的指纹匹配定位方法", 《广东工业大学学报》 *
覃玉清: "基于深度学习的WIFI定位算法", 《中国优秀硕士学位论文全文数据库》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109242097A (en) * 2018-07-12 2019-01-18 上海交通大学 The visual representation learning system and method for unsupervised learning
CN109242097B (en) * 2018-07-12 2020-09-25 上海交通大学 Visual representation learning system and method for unsupervised learning
CN108966131A (en) * 2018-07-23 2018-12-07 广州都市圈网络科技有限公司 Fusion floor estimating method based on indoor positioning
CN109040957A (en) * 2018-08-14 2018-12-18 广东小天才科技有限公司 A kind of indoor orientation method and device based on WIFI
CN109359725A (en) * 2018-10-24 2019-02-19 北京周同科技有限公司 Training method, device, equipment and the computer readable storage medium of convolutional neural networks model
CN109359725B (en) * 2018-10-24 2021-03-02 北京周同科技有限公司 Training method, device and equipment of convolutional neural network model and computer readable storage medium
CN109246608A (en) * 2018-11-16 2019-01-18 重庆小富农康农业科技服务有限公司 A kind of point-to-point localization method in interior based on WIFI location fingerprint big data analysis
CN109474887A (en) * 2018-12-05 2019-03-15 哈尔滨工业大学 Height floor differentiating method based on LTE signal
CN109474887B (en) * 2018-12-05 2020-09-01 哈尔滨工业大学 High-low floor distinguishing method based on LTE signals
CN110231593A (en) * 2019-04-11 2019-09-13 深圳市城市交通规划设计研究中心有限公司 Indoor orientation method, device, computer readable storage medium and terminal device
CN114514759A (en) * 2019-10-04 2022-05-17 三星电子株式会社 System and method for WiFi-based indoor positioning via unsupervised domain adaptation
CN111031477A (en) * 2019-12-25 2020-04-17 中山大学 Millimeter wave indoor positioning method based on deep learning

Similar Documents

Publication Publication Date Title
CN107896362A (en) A kind of WIFI location fingerprints localization method and system based on deep learning
CN107832834B (en) Method for constructing WIFI indoor positioning fingerprint database based on generation countermeasure network
CN103295198B (en) Based on redundant dictionary and the sparse non-convex compressed sensing image reconstructing method of structure
CN109272107A (en) A method of improving the number of parameters of deep layer convolutional neural networks
KR102190103B1 (en) Method of providing commercialization service of an artificial neural network
CN105846925B (en) A kind of method and device of three-dimensional MIMO OTA Channel Modelings
CN108022589A (en) Aiming field classifier training method, specimen discerning method, terminal and storage medium
CN110445581A (en) The method for reducing the channel decoding bit error rate based on convolutional neural networks
CN109191389A (en) A kind of x-ray image adaptive local Enhancement Method
CN108717178A (en) A kind of sound localization method and device based on neural network
Yin et al. Deep CSI compression for massive MIMO: A self-information model-driven neural network
CN107977949A (en) A kind of method improved based on projection dictionary to the Medical image fusion quality of study
CN109728824A (en) A kind of LDPC code iterative decoding method based on deep learning
CN111861924A (en) Cardiac magnetic resonance image data enhancement method based on evolved GAN
Klus et al. Transfer learning for convolutional indoor positioning systems
CN110222816B (en) Deep learning model establishing method, image processing method and device
Yoo et al. On the role of vit and cnn in semantic communications: Analysis and prototype validation
Chaves-Villota et al. DeepREM: Deep-Learning-Based Radio Environment Map Estimation from Sparse Measurements
CN106610486A (en) Node positioning method and device
CN112101438A (en) Left and right eye classification method, device, server and storage medium
CN105528623B (en) A kind of imaging spectrum sparse representation method based on atural object category classification redundant dictionary
CN115620831A (en) Method for generating sequence mutation fitness through loop iteration optimization and related device
US20220092391A1 (en) System and method of using neuroevolution-enhanced multi-objective optimization for mixed-precision quantization of deep neural networks
CN109800815A (en) Training method, wheat recognition methods and training system based on Random Forest model
CN110060314A (en) A kind of CT iterative approximation accelerated method and system based on artificial intelligence

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: 20180410

RJ01 Rejection of invention patent application after publication