CN104573745B - Fruit-fly classified method based on magnetic resonance imaging - Google Patents
Fruit-fly classified method based on magnetic resonance imaging Download PDFInfo
- Publication number
- CN104573745B CN104573745B CN201510030704.8A CN201510030704A CN104573745B CN 104573745 B CN104573745 B CN 104573745B CN 201510030704 A CN201510030704 A CN 201510030704A CN 104573745 B CN104573745 B CN 104573745B
- Authority
- CN
- China
- Prior art keywords
- magnetic resonance
- trypetid
- resonance imaging
- fly
- fruit
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of fruit-fly classified method based on magnetic resonance imaging, magnetic resonance imaging is carried out to trypetid to be identified using special magnetic resonance coil, according to the characteristic of MRI and simulation human vision, by the spatial domain of MRI by Fourier transformation be frequency domain, outburst area in frequency domain obtains the marking area of image space domain through inverse Fourier transform again, the data message of marking area is learnt and trained using neural network algorithm, obtain more reliable objective result, the trypetid of existing species in the objective result and knowledge data base is subjected to images match, to obtain trypetid species to be identified.Magnetic resonance imaging is used for the identification of trypetid species by the present invention, and the entomological taxonomy methods such as the outward appearance and biology of routine are not used, above method step is completed by software after completion magnetic resonance imaging, you can the species of trypetid is obtained, recognition accuracy is high.
Description
Technical field
The present invention relates to magnetic resonance imaging application field, specifically a kind of fruit-fly classified side based on magnetic resonance imaging
Method.
Background technology
Magnetic resonance imaging (Magnetic Resonance Imaging, MRI) technology be using hydrogen nuclei magnetic,
Resonance signal is produced in the presence of additional gradient magnetic and excitation, so as to detect and be depicted as the one of the structural images of interior of articles
The method of kind, is the milestone that modern physics is applied to clinical medicine domain.In addition to medical domain, MRI is used as iconography
High-end core technology develop rapidly, its hardware platform and software engineering are constantly updated, and application progressively expands.With
Device miniaturization, production domesticization and prices, progressively popularization and application, such as：Physics, chemistry, medical treatment, petrochemical industry, engage in archaeological studies, go out
All many-sides such as Passport control quarantine, the quality restriction of industrial and agricultural products, food safety detection.
In scientific research field, magnetic resonance is seldom applied to insect imaging.In the prior art, the magnetic resonance imaging of insect is not but
It is used to solve the technical problem in practical application.Trypetid easily produces harm as quarantine pest to crops, especially outer
Come the invasion of species, the phenomenon that is likely to result in amount reproduction and can not contain.Therefore, come for inspection and quarantining for import/export department
Say, rapidly recognize that trypetid species can effectively prevent the entrance of Exotic pests, improve the effect of inspection and quarantining for import/export
Rate, it appears most important.
Typically, the method mainly described by biology and outward appearance completes fruit-fly classified, and ordinary person can not also realize pair
The identification of trypetid species, it is difficult to meet inspection and quarantine requirement.Prior art also by tissue section method to trypetid dissect after again
Recognized by its internal structure, be relative complex disruptive method, and higher, the non-specialized-technical personnel of professional requirement
It is difficult to be competent at, is not easy to the practical operation of inspection and quarantining for import/export department.In the prior art, although magnetic resonance can to trypetid into
Picture, but can not be identified and recognize to trypetid according to MRI, it is difficult to solve this real technical problem.
The content of the invention
In view of this, the technical problem to be solved in the present invention is to provide one kind and can realized using MRI to reality
The trypetid recognition methods based on magnetic resonance imaging of fly identification.
The technical solution of the present invention is to provide the trypetid recognition methods based on magnetic resonance imaging of following steps, bag
Include following steps：
1) in magnetic resonance imaging system, trypetid to be sorted is imaged using small size cylinder type magnetic resonance radio-frequency coil,
Trypetid is placed in radio-frequency coil during imaging, trypetid MRI is obtained；The magnetic resonance imaging system main field field strength 1.0
More than tesla, imaging region field strength inhomogeneities is less than 2ppm.
2) according to obtained trypetid MRI, conspicuousness detection is carried out to image, become by trying to achieve image Fourier
The residual error of amplitude spectrum after changing, the saliency map picture of spatial domain is obtained through inverse Fourier transform, can be shown from saliency map picture
Write region；
3) the corresponding data message of collection marking area, using the neural network algorithm of extreme learning machine to the notable area
The data message in domain carries out online Fast Learning in real time, obtains objective result；
4) objective result and knowledge data base are subjected to images match, described knowledge data base contains classification
The characteristic of a variety of trypetids, described characteristic is also obtained by magnetic resonance imaging；
5) according to the result of matching, the species whether trypetid to be identified belongs in knowledge data base is judged, if then
It is determined which the species belonged in knowledge data base.
Using the method for the present invention, compared with prior art, the present invention has advantages below：The present invention uses special magnetic
Resonance coil carries out magnetic resonance imaging to trypetid to be identified, according to the characteristic of MRI and simulation human vision, by magnetic
The spatial domain of resonance image is frequency domain by Fourier transformation, and the outburst area in frequency domain is obtained through inverse Fourier transform again
The marking area of image space domain, is learnt and is trained to the data message of marking area using neural network algorithm, obtained
More reliable objective result, by the trypetid of species carries out images match in the objective result and knowledge data base, with
To trypetid species to be identified.Magnetic resonance imaging is used for the identification of trypetid species by the present invention, without carrying out biology and anatomy
Analysis contrast, complete to pass through software after magnetic resonance imaging completing above method step, you can obtain the species of trypetid, identification is accurate
Exactness is high.
As an improvement, step 2) in the saliency map picture by average filter carry out it is once smooth, to eliminate minority
Isolated significant point.In view of the local group effect (present invention is based on the simulation to human vision) of human eye vision, average
More preferable visual effect can be obtained after filter filtering.
As an improvement, completing step 3 using integrated neural network), concretely comprise the following steps：It is right respectively by multiple component classifiers
Corresponding training dataset has the repeated sampling put back to, parallel individually training, then uses majority voting method to carry out integrated, obtains
Objective result after combining classifiers, is used as step 4) matching object.Using the step, integrated neural network performance is better than single
Individual neutral net is made that explanation, and neutral net number is sufficiently large in theory, then error levels off to 0.
As an improvement, step 4) described in images match use the image matching method based on Scale invariant features transform
Carry out, be specially：Picture position on a, all yardsticks of search, recognized by gaussian derivative function it is potential for yardstick and
The point of interest of invariable rotary；B, on the point of interest of each candidate, position and chi are determined by a fine model of fitting
Degree；C, the gradient direction based on image local, distribute to the one or more directions of each point of interest location, it is all behind to figure
As the operation of data both relative to the direction of point of interest, yardstick and position enters line translation, so as to provide for these conversion not
Denaturation；D, in the neighborhood around each key point, the gradient of image local is measured on selected yardstick.Using above-mentioned matching
Method, with to rotation, scaling and brightness change maintain the invariance, to visual angle change, affine transformation and noise keep compared with
The stability of high level.
Brief description of the drawings
Fig. 1 is the FB(flow block) of fruit-fly classified method of the invention based on magnetic resonance imaging.
Embodiment
The invention will be further described with specific embodiment below in conjunction with the accompanying drawings, but the present invention is not restricted to these
Embodiment.
The present invention covers any replacement, modification, equivalent method and scheme made in the spirit and scope of the present invention.For
Make the public have the present invention thoroughly to understand, concrete details is described in detail in present invention below preferred embodiment, and
Description without these details can also understand the present invention completely for a person skilled in the art.In addition, the accompanying drawing of the present invention
In in order to illustrate the need for, be not drawn to scale accurately completely, be explained herein.
As shown in figure 1, the trypetid recognition methods based on magnetic resonance imaging of the present invention, including implementation steps in detail below：
1) in magnetic resonance imaging system, trypetid to be sorted is imaged using small size cylinder type magnetic resonance radio-frequency coil,
Trypetid is placed in radio-frequency coil during imaging, trypetid MRI is obtained；The magnetic resonance imaging system main field field strength 1.0
More than tesla, imaging region field strength inhomogeneities is less than 2ppm.
2) according to obtained trypetid MRI, conspicuousness detection is carried out to image, become by trying to achieve image Fourier
The residual error of amplitude spectrum after changing, the saliency map picture of spatial domain is obtained through inverse Fourier transform, can be shown from saliency map picture
Write region；
3) the corresponding data message of collection marking area, using the neural network algorithm of extreme learning machine to the notable area
The data message in domain carries out online Fast Learning in real time, obtains objective result；
4) objective result and knowledge data base are subjected to images match, described knowledge data base contains classification
The characteristic of a variety of trypetids, described characteristic is also obtained by magnetic resonance imaging；
5) according to the result of matching, the species whether trypetid to be identified belongs in knowledge data base is judged, if then
It is determined which the species belonged in knowledge data base.
Step 2) in the saliency map picture carried out by average filter it is once smooth, it is a small number of isolated notable to eliminate
Point.
Step 3 is completed using integrated neural network), concretely comprise the following steps：By multiple component classifiers respectively to corresponding instruction
Practicing data set has the repeated sampling put back to, parallel individually training, then uses majority voting method to carry out integrated, obtains grader collection
Objective result after, is used as step 4) matching object.
Step 4) described in images match using based on Scale invariant features transform image matching method carry out, specifically
For：Picture position on a, all yardsticks of search, is recognized potential for yardstick and invariable rotary by gaussian derivative function
Point of interest；B, on the point of interest of each candidate, position and yardstick are determined by a fine model of fitting；C, it is based on
The gradient direction of image local, distributes to the one or more directions of each point of interest location, it is all behind to view data
Operation enters line translation both relative to the direction of point of interest, yardstick and position, so as to provide the consistency for these conversion；D,
In neighborhood around each key point, the gradient of image local is measured on selected yardstick.
The mode of present invention simulation human eye processing visual information, proposition one kind " conspicuousness detection+neural network model+
The MRI image target identification framework of SIFT feature matching ".First, the conspicuousness area in image is positioned using vision noticing mechanism
Domain, is gathered by a few sample to salient region, online to build pixel classifications model；Finally, the mesh that image segmentation is obtained
The trypetid MRI image marked with known class in MRI image data storehouse (knowledge data base) compares, and matches and realizes through SIFT feature
Target classification.
1st, conspicuousness is detected.There is each second substantial amounts of visual information to enter human eye.Mechanism if none of wisdom is come
The extraneous data in vision is filtered out, processing total data will be a God-awful thing in real time.High-level is cognitive and multiple
Reason is lived together, such as object is cognitive or scene understands, be dependent on these by converted tractable data.This mechanism is just
It is vision attention, it is therefore desirable to which the salient region in image is identified.
Model based on frequency-domain analysis mainly uses the methods such as Fourier transformation that image is transformed from a spatial domain into frequency domain,
And frequency domain information is analyzed and processed, notable feature is searched out, then contravariant changes to spatial domain and obtains saliency map.By trying to achieve image
The residual error of amplitude spectrum after Fourier transformation, direct detection image marking area is carried out through inverse Fourier transform.With other conspicuousnesses
Comparison between detecting methods, compose the calculating speed of residual error method quickly.More specifically conspicuousness detecting step is as follows：
For given MRI I (x), two dimensional discrete Fourier transform F [I (x)] is carried out to it first, by image
Frequency domain is changed to by transform of spatial domain, amplitude A (f) and phase P (f) information is obtained：
A (f)=| F [I (x)] | (1)
Then amplitude is taken the logarithm, obtains log spectrum L (f)：
L (f)=log (A (f)) (3)
F represents two dimensional discrete Fourier transform in formula, | | amplitude computing is represented,Represent phase operation；Because log is bent
Line meets local linear condition, so with local average wave filter hn(f) carry out smoothly obtaining the general shape of log spectrums to it：
V (f)=L (f) * hn(f) (4)
Wherein hn(f) be a n × n matrix, be defined as follows：
It is then the description to the sudden change region in image to compose residual error R (f)：
R (f)=L (f)-V (f) (6)
By inverse Fourier transform, saliency map picture can be obtained in spatial domain.
S (x)=| F-1[exp{R(f)+jP(f)}]|2 (7)
The value of every represents the significance of the position on saliency map.In view of the local group effect of human eye vision, it is
A small number of isolated significant points are eliminated, more preferable visual effect are obtained, we are carried out with average filter again after S (x) is obtained
It is once smooth, obtain final saliency map Z (x).
Z (x)=S (x) * hn(f) (8)
It is a kind of batch processing to compose operation of the residual error algorithm to pixel, and algorithm is simple, quick, easily realize, and has for noise
There is certain robustness.
2nd, online Fast Learning in real time.It is using the purpose of machine learning：1. directly simulated using neural network algorithm
The mankind " brain-eye " nervous system；2. unstructured information is converted to by " study " can computation model.
The present invention uses a kind of new neutral net of batch processing training data --- extreme learning machine (Extreme
Learning machine, ELM) it is used as the basis of machine learning algorithm.ELM is a kind of Single hidden layer feedforward neural networks
(Single-hidden layer feedforward networks, SLFNs).The training sample set given to oneThe SLFNs for having L hidden layer node is expressed as：
Wherein aiAnd biIt is the parameter of hidden node, can randomly generating independently of training data.K(ai, bi, xj) it is i-th
Hidden node corresponds to the output item of input.βiIt is connection weight of i-th of hidden node to output node.If given training sample
This is, it is known that and aiAnd biIt has been randomly generated that, then K (ai, bi, xj) can calculate, formula (9) turns into a linear system, wherein only
There is βiIt is the unknown, can be solved by linear algebra approach.
Based on above-mentioned principle, under given training set, the performance of single hidden layer Feedback Neural Network is completely by its hidden layer node
Determined with the connection weight of output node, and it is unrelated with connection weight, the deviant of hidden layer node etc. with input.Thus, it is possible to
Mathematical measure solves the analytic solutions rather than iterative approximation solution of crucial connection weight, so that ELM algorithms are substantially optimal, it is to avoid
There is the situation of local minimum in neutral net based on gradient descent method iterative.
Output result is obtained using K simple ELM, then the posterior probability of each sample is obtained by integrated approach, is connect
Get off and sample class is calculated according to posterior probability.This method effectively solves the unstability of single ELM study, and by
In using integrated method, ELM Generalization Capability is improved.
Integrated neural network performance can be better than single Neural：Assuming that integrated classifier is by N number of incoherent
Neutral net is constituted, and the error in classification of each neutral net is p, and the error of each neutral net is separate, using many
Ballot method is counted, then integrated error is
From formula (10), as p ＜ 0.5, E with N increase monotone decreasing.So, if each member's neutral net
Classification accuracy rate is all higher than 50%, and the error of each member network is separate, then when member's neutral net number N is enough
When big, Ensemble classifier error is intended to 0.Although in practical application, member network's number be it is limited multiple, each member network's
Error generally can not obtain 100% Ensemble classifier accuracy nor separate, but can generally obtain higher than single
The nicety of grading of individual neutral net.
Bagging (Bootstrap are relied on because ELM models are not highly stable graders, therefore in present invention improvement
Aggregating) algorithm：There is the repeated sampling (Bootstrap put back to training set respectively by multiple component classifiers
Sampling), parallel individually training, is then carried out integrated using ballot method.
3rd, the images match based on Scale invariant features transform (SIFT) feature
The essence of SIFT algorithms is that key point (point of interest) is searched on different metric spaces, and calculates key point
Direction.The key point that SIFT is found will not change because of illumination, the factor such as affine transformation and noise, such as angle point, marginal point, dark
The bright spot in area and the dim spot in clear zone etc..SIFT algorithms are decomposed into following four step：
1) metric space extremum extracting：The picture position searched on all yardsticks.Recognized by gaussian derivative function latent
For yardstick and the point of interest of invariable rotary.
2) crucial point location：On the position of each candidate, position and chi are determined by the fine model of a fitting
Degree.The selection gist of key point is in their degree of stability.
3) direction is determined：Gradient direction based on image local, distributes to each one or more directions in key point position.
Operation to view data behind all enters line translation both relative to the direction of key point, yardstick and position, so as to provide pair
In the consistency of these conversion.
4) key point is described：In the neighborhood around each key point, the ladder of image local is measured on selected yardstick
Degree.These gradients are transformed into a kind of expression, and this expression allows the deformation and illumination variation than larger local shape.
The characteristics of SIFT algorithms：
1) SIFT feature is the local feature of image, its to rotation, scaling, brightness change maintains the invariance, to regarding
Angle change, affine transformation, noise also keep a certain degree of stability；
2) unique (Distinctiveness) is good, informative, it is adaptable to carried out in magnanimity property data base fast
Speed, accurately matching；
3) volume, substantial amounts of SIFT feature vector can also be produced even if a small number of several objects；
4) high speed, optimized SIFT matching algorithms even can reach real-time requirement；
5) scalability, very easily can be combined with the characteristic vector of other forms.
Only preferred embodiments of the present invention are described above, but are not to be construed as limiting the scope of the invention.This
Invention is not only limited to above example, and its concrete structure allows to change.In a word, all guarantors in independent claims of the present invention
The various change made in the range of shield is within the scope of the present invention.
Claims (4)
1. a kind of fruit-fly classified method based on magnetic resonance imaging, it is characterised in that：Including following steps：
1) in magnetic resonance imaging system, trypetid to be sorted is imaged using small size cylinder type magnetic resonance radio-frequency coil, is imaged
When trypetid is placed in radio-frequency coil, obtain trypetid MRI；The magnetic resonance imaging system main field field strength 1.0 it is special this
More than drawing, imaging region field strength inhomogeneities is less than 2ppm；
2) according to obtained trypetid MRI, conspicuousness detection is carried out to image, by trying to achieve after image Fourier transformation
Amplitude spectrum residual error, the saliency map picture of spatial domain is obtained through inverse Fourier transform, notable area is can obtain from saliency map picture
Domain；
3) the corresponding data message of collection marking area, using the neural network algorithm of extreme learning machine to the marking area
Data message carries out online Fast Learning in real time, obtains objective result；
4) objective result and knowledge data base are subjected to images match, described knowledge data base contains a variety of of known class
The characteristic of trypetid, described characteristic is also obtained by magnetic resonance imaging；
5) according to the result of matching, the Known Species whether trypetid to be sorted belongs in knowledge data base are judged, if then determining
Which the specific species belonged in knowledge data base.
2. the fruit-fly classified method according to claim 1 based on magnetic resonance imaging, it is characterised in that：Using integrated nerve
Network completes step 3), concretely comprise the following steps：There is the repetition put back to corresponding training dataset respectively by multiple component classifiers
Then sampling, parallel individually training uses majority voting method to carry out integrated, obtains the objective result after combining classifiers, as
Step 4) matching object.
3. the fruit-fly classified method according to claim 2 based on magnetic resonance imaging, it is characterised in that：Step 2) in institute
Saliency map picture is stated once smooth by average filter progress, to eliminate a small number of isolated significant points.
4. the fruit-fly classified method according to claim 3 based on magnetic resonance imaging, it is characterised in that：Step 4) described in
Images match using based on Scale invariant features transform image matching method carry out, be specially：On a, all yardsticks of search
Picture position, recognized by gaussian derivative function potential for yardstick and the point of interest of invariable rotary；B, it is each wait
On the point of interest of choosing, position and yardstick are determined by the fine model of a fitting；C, the gradient direction based on image local,
Distribute to the one or more directions of each point of interest location, it is all behind the operation to view data both relative to point of interest
Line translation is entered in direction, yardstick and position, so as to provide the consistency for these conversion；D, the neighborhood around each key point
It is interior, the gradient of image local is measured on selected yardstick.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510030704.8A CN104573745B (en) | 2015-01-21 | 2015-01-21 | Fruit-fly classified method based on magnetic resonance imaging |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510030704.8A CN104573745B (en) | 2015-01-21 | 2015-01-21 | Fruit-fly classified method based on magnetic resonance imaging |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104573745A CN104573745A (en) | 2015-04-29 |
| CN104573745B true CN104573745B (en) | 2017-10-03 |
Family
ID=53089764
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510030704.8A Expired - Fee Related CN104573745B (en) | 2015-01-21 | 2015-01-21 | Fruit-fly classified method based on magnetic resonance imaging |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104573745B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573699B (en) * | 2015-01-21 | 2017-11-24 | 中国计量学院 | Trypetid recognition methods based on middle equifield intensity magnetic resonance anatomy imaging |
| CN104933724B (en) * | 2015-06-25 | 2019-07-26 | 中国计量学院 | The Automatic image segmentation method of trypetid magnetic resonance image |
| CN106577350B (en) * | 2016-11-22 | 2020-10-09 | 深圳市沃特沃德股份有限公司 | Pet type identification method and device |
| CN112101396A (en) * | 2019-06-18 | 2020-12-18 | 深圳先进技术研究院 | Classification method, classification device, classification equipment and storage medium |
| CN111291702A (en) * | 2020-02-20 | 2020-06-16 | 北京嘉景生物科技有限责任公司 | Identification method for distinguishing locust species by image recognition technology |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177266A (en) * | 2013-04-07 | 2013-06-26 | 青岛科技大学 | Intelligent stock pest identification system |
| CN104573699A (en) * | 2015-01-21 | 2015-04-29 | 中国计量学院 | Trypetid identification method based on medium field intensity magnetic resonance dissection imaging |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7496228B2 (en) * | 2003-06-13 | 2009-02-24 | Landwehr Val R | Method and system for detecting and classifying objects in images, such as insects and other arthropods |
-
2015
- 2015-01-21 CN CN201510030704.8A patent/CN104573745B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177266A (en) * | 2013-04-07 | 2013-06-26 | 青岛科技大学 | Intelligent stock pest identification system |
| CN104573699A (en) * | 2015-01-21 | 2015-04-29 | 中国计量学院 | Trypetid identification method based on medium field intensity magnetic resonance dissection imaging |
Non-Patent Citations (1)
| Title |
|---|
| 基于目标场法的磁共振成像小尺寸射频线圈设计;徐文龙;《高电压技术》;20141231;第40卷(第12期);3786-3792 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104573745A (en) | 2015-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kesav et al. | Efficient and low complex architecture for detection and classification of Brain Tumor using RCNN with Two Channel CNN | |
| Li et al. | Crop pest recognition in natural scenes using convolutional neural networks | |
| CN104573699B (en) | Trypetid recognition methods based on middle equifield intensity magnetic resonance anatomy imaging | |
| Zhang et al. | Image based fruit category classification by 13-layer deep convolutional neural network and data augmentation | |
| Li et al. | Robust capsule network based on maximum correntropy criterion for hyperspectral image classification | |
| Pitchai et al. | Brain tumor segmentation using deep learning and fuzzy K-means clustering for magnetic resonance images | |
| Liu et al. | Spatial-spectral kernel sparse representation for hyperspectral image classification | |
| Unnikrishnan et al. | Deep learning architectures for land cover classification using red and near-infrared satellite images | |
| Li et al. | Deep convolutional neural networks for detecting secondary structures in protein density maps from cryo-electron microscopy | |
| Lu et al. | Classification of Camellia (Theaceae) species using leaf architecture variations and pattern recognition techniques | |
| CN104573745B (en) | Fruit-fly classified method based on magnetic resonance imaging | |
| Rahaman et al. | An efficient multilevel thresholding based satellite image segmentation approach using a new adaptive cuckoo search algorithm | |
| Romaguera et al. | Left ventricle segmentation in cardiac MRI images using fully convolutional neural networks | |
| Ghesu et al. | Marginal space deep learning: Efficient architecture for detection in volumetric image data | |
| Yan et al. | Brain tissue segmentation based on graph convolutional networks | |
| Taghipour et al. | Hyperspectral anomaly detection using spectral–spatial features based on the human visual system | |
| CN104573746B (en) | Trypetid category identification method based on magnetic resonance imaging | |
| Mata et al. | Automated neuron detection in high-content fluorescence microscopy images using machine learning | |
| Abisha et al. | Brinjal leaf diseases detection based on discrete Shearlet transform and Deep Convolutional Neural Network | |
| CN104933724A (en) | Automatic image segmentation method of trypetid magnetic resonance image | |
| Bellary | The MRI knee pain classification using CNN algorithm and segmentation using clustering algorithm | |
| Kumari et al. | A genetic algorithm optimized artificial neural network for the segmentation of MR images in frontotemporal dementia | |
| Wan et al. | Spectral-spatial hyperspectral image classification combining multi-scale bi-exponential edge-preserving filtering and susan edge detector | |
| Lu et al. | A novel key features screening method based on extreme learning machine for Alzheimer’s disease study | |
| Pan et al. | Leukocyte image segmentation using novel saliency detection based on positive feedback of visual perception |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171003 Termination date: 20180121 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |