CN109190641A - A kind of cervical cell feature extracting method based on LDA and transfer learning - Google Patents
A kind of cervical cell feature extracting method based on LDA and transfer learning Download PDFInfo
- Publication number
- CN109190641A CN109190641A CN201810993450.3A CN201810993450A CN109190641A CN 109190641 A CN109190641 A CN 109190641A CN 201810993450 A CN201810993450 A CN 201810993450A CN 109190641 A CN109190641 A CN 109190641A
- Authority
- CN
- China
- Prior art keywords
- feature
- lda
- cervical cell
- extracting method
- method based
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—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]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of cervical cell feature extracting method based on LDA and transfer learning, it is characterised in that: the feature extracting method based on deep learning network model, first with one deep learning network model of large data sets training and preservation model;Then transfer learning is utilized, cervical cell data set is input in trained deep learning network model, primitive character collection is extracted;Dimension-reduction treatment finally is carried out to primitive character collection using LDA, obtains final feature set, this feature collection covers the more information of data set, improves the accuracy rate of data classification.
Description
Technical field
It is specifically a kind of to utilize transfer learning the present invention relates to feature extraction, deep learning, transfer learning and Data Dimensionality Reduction
Trained deep learning network model is come the method for extracting feature.
Background technique
With the development of the subjects such as network knowledge tissue, artificial intelligence, the technologies such as image recognition, speech recognition increasingly by
To extensive concern.Identification technology is sought to a kind of research object, is identified and is classified according to its certain feature.It can recognize
To realize that identification and classification to research object need first to extract the feature of research object.However, with the expansion of practical activity
Greatly, deeply with the needs of socialization, people not only need to identify many things of classification number, but also identified contents of object
It becomes increasingly complex.Especially because the raising of scientific and technological level, so that a variety of different research objects " image conversion " or " number
Change ", this is more demanding to Feature Extraction Technology, and the feature for needing to extract can cover more legacy data information, also wants
Calculation amount is reduced, runing time is reduced.
Existing Feature Extraction Technology is all artificially to find out feature different between data set and extract respectively, is not only covered
Raw information it is few, and each feature artificially will find and calculate, and subjectivity is strong, extracts with this method
Accuracy rate of the feature set in identification be not very high;Feature extracting method based on deep learning network model is to utilize machine
The feature of learning data set, the raw information that this feature collection is covered is more, more simplifies and ability to express is stronger, uterine neck can be improved
Cell image recognition rate.
Summary of the invention
The purpose of the present invention is to provide a kind of cervical cell feature extracting method based on LDA and transfer learning, with solution
Certainly the problems mentioned above in the background art.
To achieve the above object, the invention provides the following technical scheme: a kind of cervical cell based on LDA and transfer learning
Feature extracting method, comprising the following steps:
(1), training deep learning network model: deep learning network model is first trained with existing large data sets, and is protected
Deposit model;
(2), extract feature using transfer learning: the model that transfer learning saves inputs cervical cell data set, and extracts
Feature;
(3), LDA carries out dimensionality reduction to primitive character collection: carrying out dimension-reduction treatment to primitive character collection using LDA, makes feature set
It more simplifies, improves Classification and Identification rate.
Compared with prior art, the beneficial effects of the present invention are: using machine learning data set feature, this feature collection contain
The raw information of lid is more, more simplifies and ability to express is stronger, help quickly and accurately to carry out the identification of research object with
Classification has good robustness, accuracy and low complex degree, and high to the Classification and Identification rate of cervical cell.
Detailed description of the invention
Fig. 1 is flow chart of the invention;
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Specific step is as follows for feature extracting method based on deep learning network model:
A, deep learning network model: deep learning network model needs very big data set that can just train, so
Using existing large data sets training pattern, make model that can accurately extract articulate feature;
B, transfer learning: retain the parameter of all convolutional layers in trained deep learning network model, only replace most
The full articulamentum of later layer, the output node vector of the last layer can be by the feature as any research object data set after modification
Vector.
Deep learning network model is to extract feature using every layer of convolutional layer, does not need artificially to be arranged and extracts certain spy
Sign, enhanced convenience;Transfer learning makes trained model in a problem by simply adjustment new it is suitable for one
The problem of, solve the problems, such as big data mark and training time.
As shown in Figure 1, a kind of cervical cell feature extracting method based on LDA and transfer learning, comprising the following steps:
A, training deep learning network model: being inputted with existing large data sets and first train deep learning network model,
And preservation model, make model that can accurately extract articulate feature;
B, extract feature using transfer learning: the model that transfer learning saves inputs cervical cell data set, and extracts spy
Sign;
C, LDA carries out dimensionality reduction to primitive character collection: carrying out dimension-reduction treatment to primitive character collection using LDA, makes feature set more
Add and simplify, is conducive to improve cervical cell Classification and Identification rate.
In conclusion feature extracting method provided by the invention utilizes the feature of machine learning data set, this feature collection is contained
The raw information of lid is more, more simplifies and ability to express is stronger, help quickly and accurately to carry out the identification of research object with
Classification has good robustness, accuracy and low complex degree, and high to the Classification and Identification rate of cervical cell.
It although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of variations, modification, replacement can be carried out to these embodiments without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is defined by the appended.
Claims (1)
1. a kind of cervical cell feature extracting method based on LDA and transfer learning, it is characterised in that: the following steps are included:
A, it trains deep learning network model and saves;
B, the deep learning network model that transfer learning has saved;
C, the cervical cell data set of oneself is input in model and extracts feature;
D, dimension-reduction treatment is carried out to characteristic using LDA and obtains final feature set.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810993450.3A CN109190641A (en) | 2018-08-29 | 2018-08-29 | A kind of cervical cell feature extracting method based on LDA and transfer learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810993450.3A CN109190641A (en) | 2018-08-29 | 2018-08-29 | A kind of cervical cell feature extracting method based on LDA and transfer learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109190641A true CN109190641A (en) | 2019-01-11 |
Family
ID=64916793
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810993450.3A Pending CN109190641A (en) | 2018-08-29 | 2018-08-29 | A kind of cervical cell feature extracting method based on LDA and transfer learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109190641A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106971174A (en) * | 2017-04-24 | 2017-07-21 | 华南理工大学 | A kind of CNN models, CNN training methods and the vein identification method based on CNN |
CN107391599A (en) * | 2017-06-30 | 2017-11-24 | 中原智慧城市设计研究院有限公司 | Image search method based on style and features |
US20180189615A1 (en) * | 2017-01-03 | 2018-07-05 | Samsung Electronics Co., Ltd. | Electronic apparatus and method of operating the same |
CN108281183A (en) * | 2018-01-30 | 2018-07-13 | 重庆大学 | Cervical smear image diagnostic system based on convolutional neural networks and transfer learning |
CN108416379A (en) * | 2018-03-01 | 2018-08-17 | 北京羽医甘蓝信息技术有限公司 | Method and apparatus for handling cervical cell image |
-
2018
- 2018-08-29 CN CN201810993450.3A patent/CN109190641A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180189615A1 (en) * | 2017-01-03 | 2018-07-05 | Samsung Electronics Co., Ltd. | Electronic apparatus and method of operating the same |
CN106971174A (en) * | 2017-04-24 | 2017-07-21 | 华南理工大学 | A kind of CNN models, CNN training methods and the vein identification method based on CNN |
CN107391599A (en) * | 2017-06-30 | 2017-11-24 | 中原智慧城市设计研究院有限公司 | Image search method based on style and features |
CN108281183A (en) * | 2018-01-30 | 2018-07-13 | 重庆大学 | Cervical smear image diagnostic system based on convolutional neural networks and transfer learning |
CN108416379A (en) * | 2018-03-01 | 2018-08-17 | 北京羽医甘蓝信息技术有限公司 | Method and apparatus for handling cervical cell image |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107333071A (en) | Video processing method and device, electronic equipment and storage medium | |
CN103035135B (en) | Children cognitive system based on augment reality technology and cognitive method | |
CN104463101B (en) | Answer recognition methods and system for character property examination question | |
CN106778584A (en) | A kind of face age estimation method based on further feature Yu shallow-layer Fusion Features | |
CN106372581A (en) | Method for constructing and training human face identification feature extraction network | |
WO2023185785A1 (en) | Image processing method, model training method, and related apparatuses | |
CN104636757B (en) | A kind of food image recognition methods based on deep learning | |
CN108984745A (en) | A kind of neural network file classification method merging more knowledge mappings | |
CN106295584A (en) | Depth migration study is in the recognition methods of crowd's attribute | |
CN110534132A (en) | A kind of speech-emotion recognition method of the parallel-convolution Recognition with Recurrent Neural Network based on chromatogram characteristic | |
CN107092894A (en) | A kind of motor behavior recognition methods based on LSTM models | |
CN108197294A (en) | A kind of text automatic generation method based on deep learning | |
CN106548159A (en) | Reticulate pattern facial image recognition method and device based on full convolutional neural networks | |
CN107341463A (en) | A kind of face characteristic recognition methods of combination image quality analysis and metric learning | |
CN109255296A (en) | A kind of daily Human bodys' response method based on depth convolutional neural networks | |
CN107463954A (en) | A kind of template matches recognition methods for obscuring different spectrogram picture | |
CN103714331A (en) | Facial expression feature extraction method based on point distribution model | |
CN112906550B (en) | Static gesture recognition method based on watershed transformation | |
CN108829810A (en) | File classification method towards healthy public sentiment | |
CN110363233A (en) | A kind of the fine granularity image-recognizing method and system of the convolutional neural networks based on block detector and Fusion Features | |
CN105117740A (en) | Font identification method and device | |
CN108509833A (en) | A kind of face identification method, device and equipment based on structured analysis dictionary | |
CN109784159A (en) | The processing method of scene image, apparatus and system | |
CN110046544A (en) | Digital gesture identification method based on convolutional neural networks | |
CN105550705A (en) | Electroencephalogram (EEG) signal recognition method based on improved self-training learning |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190111 |
|
WD01 | Invention patent application deemed withdrawn after publication |