CN104834943A - Brain tumor classification method based on deep learning - Google Patents
Brain tumor classification method based on deep learning Download PDFInfo
- Publication number
- CN104834943A CN104834943A CN201510268230.0A CN201510268230A CN104834943A CN 104834943 A CN104834943 A CN 104834943A CN 201510268230 A CN201510268230 A CN 201510268230A CN 104834943 A CN104834943 A CN 104834943A
- Authority
- CN
- China
- Prior art keywords
- brain tumor
- degree
- deep learning
- input
- parameter
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Abstract
The invention discloses a brain tumor classification method based on deep learning, and the method comprises the steps: employing Gabor wavelet transform to extract the textural features of a brain tumor firstly during the extracting of brain tumor features; building a deep learning method according to stacking noise-reduction automatic coding, and then employing the deep learning to extract higher-layer features from the textural features; employing a concentric circle method to extract the shape features of the brain tumor secondly, and combining the shape features with the higher-layer features extracted through the deep learning to form an augmented feature vector; enabling the features to serve as input placed in a support vector machine, and carrying out training to obtain a classifier; finally employing the same method to extract a feature vector for a test sample, and employing the classifier obtained through training to carry out classification. The method improves the diagnosis accuracy of a doctor, and provides usable information for the making of an operation scheme for the brain tumor.
Description
Technical field
The invention belongs to computer-aided diagnosis field, more specifically say, relate to a kind of brain tumor sorting technique based on degree of depth study.
Background technology
Degree of depth study is the popular domain of current machine Learning Studies, and these researchs relate to the every aspect of mankind's activity, as speech recognition, and Face datection, target following, semantic parsing etc.But degree of depth study at present is also seldom applied to computer-aided diagnosis field.
At present, what computer-aided diagnosis research was comparatively ripe is in mammary gland and pulmonary lesion, research proves, computer-aided diagnosis system can be supplied to the objective computer diagnosis result of doctor one as the second reference, fails to pinpoint a disease in diagnosis serve positive effect for accuracy rate, the minimizing improving diagnosis.
But the research of computer-aided diagnosis system in brain tumor is also in the starting stage, and large quantifier elimination is in progress.But brain tumor to be united modal disease except cerebrovascular disease ectoneural system, be considered to one of refractory neoplasm the most thorny in neurosurgical treatment, seriously threaten the healthy of the mankind.Add up according to domestic epidemiologic data, China's brain tumor morbidity rate is 32,/10 ten thousand population, and the incidence rate in an average annual is 4 ~ 10,/10 ten thousand populations (i.e. annual neopathy 4 ~ 130,000 people in China 1,300,000,000 population).Brain tumor accounts for 1% ~ 6% of general tumour, in the death that whole body malignant tumour causes, account for 2%, occupies the tenth.
The early detection of brain tumor can improve cure rate greatly, and the differentiation of benign from malignant tumors has a great impact the therapeutic scheme formulating patient.If more reliably can judge the good pernicious misery that just may alleviate patient of tumour in the preoperative according to patient symptom, even can cure some patients, and the opportunity for the treatment of further can be won, likely improve the life quality of patient and extend life cycle.
The good pernicious method of existing judgement brain tumor mainly first extracts the textural characteristics of brain tumor, and recycling PCA dimensionality reduction, finally utilizes support vector machine to classify.But this method is confined to single low-level feature.And the maximum effect of degree of depth study, be apish thinking, from low-level feature, take out the high-level characteristic having more discernment.Document (Suk H I, Shen D.Deep learning-based feature representationfor AD/MCI classification [M] //Medical Image Computing and Computer-AssistedIntervention – MICCAI 2013.Springer Berlin Heidelberg, 2013:583-590.) successfully utilize the degree of depth to learn, from the grey matter volume of 93 area-of-interests, extract high-level characteristic, improve the classification accuracy of AD/MCI.
Summary of the invention
The object of the invention is to the deficiency overcoming existing brain tumor sorting technique, a kind of good pernicious sorting technique of brain tumor based on degree of depth study is provided, find the potential character representation of high level on the low-level feature such as texture, shape basis, improve the discernment of feature, reduce the error rate of classification, assist physician judges the type of tumour more accurate and visually, formulates appropriate operation plan.
The technical solution adopted for the present invention to solve the technical problems is as follows:
Based on the brain tumor sorting technique of degree of depth study, its flow process as shown in Figure 1, specifically comprises the following steps:
Step one: the MRI image browsing the patient suffering from brain tumor, chooses one deck that its midbrain tumors is maximum, carries out the pre-service such as denoising, enhancing, and brain tumor split it;
Step 2: the textural characteristics extracting brain tumor based on Gabor wavelet, extracts the Gabor wavelet textural characteristics on different scale and direction;
Step 3: for each tumour constructs the concentric circles of one group of k radius equivalent increase, to each radius increment, calculate the ratio of the part of this circle and tumour overlap, form the k dimensional feature vector of each tumour;
Step 4: setting entire deep learning model is the degree of deep learning network comprising L hidden layer, using the texture feature vector in step 2 as input, the degree of depth is utilized to learn to carry out the extraction of high-level characteristic, first pre-training is successively carried out, utilize noise reduction automatic coding, coding and decoding is carried out to input, adjustment parameter between layers, error after making original input and rebuilding is minimum, the output of first hidden layer is obtained by this parameter, as the input of second hidden layer, by that analogy, until parameter after obtaining L hidden layer degree of deep learning process;
Step 5: finally increase an output layer at the network of pre-training, by there being the data of label, utilizing backpropagation and gradient descent method to finely tune the parameter of whole network from top to bottom, obtaining final argument;
Step 6: with the output of network as the high-level characteristic extracted, be combined with low layer textural characteristics and shape facility and form augmented feature vector, with Data Dimensionality Reduction Algorithm PCA, dimensionality reduction calculating is carried out to this proper vector, obtain the final feature of all samples;
Step 7: by the proper vector input support vector machine training classifier after dimensionality reduction;
Step 8: extract same feature according to feature extracting method identical before to test data, obtains the proper vector of all test datas, is entered in sorter, obtains classification results, statistics accuracy rate.
It should be noted that:
For different image data sets, the setting of the model parameter of degree of depth study often has a great difference, needs to be determined by experiment best parameter.
The invention has the beneficial effects as follows:
The framework that traditional brain tumor sorting technique based on textural characteristics and the degree of depth learn combines by the present invention, propose one and can extract more higher level of abstraction feature from low layer textural characteristics, and improve the transaction module of the good pernicious classification accuracy of brain tumor accordingly; The present invention has the following advantages:
1, compared with traditional brain tumor sorting technique, the present invention not only considers the textural characteristics distinguishing brain tumor, also according to the characteristic of innocent and malignant tumour, have selected a kind of method of altering course with a concentric circle carried for brain tumor shape, for extracting the shape facility of brain tumor, improve brain tumor discernment further;
2, compared with traditional degree of deep learning framework, the present invention inputs with the multidimensional texture feature extracted, and feature is extracted from the gray-scale value of each pixel unlike the study of the former degree of depth, on the basis of low layer textural characteristics, more abstract high-level characteristic is obtained with the framework of degree of depth study, to combine formation augmented feature vector with original textural characteristics, make the characteristics of image of acquisition have better resolution.
Accompanying drawing explanation
Fig. 1 is the brain tumor sorting technique process flow diagram based on degree of depth study provided by the invention;
Fig. 2 is the theory diagram of degree of depth study pre-training;
Fig. 3 is the theory diagram of degree of depth study fine setting.
Specific implementation
The brain tumor sorting technique based on degree of depth study that the present invention proposes specifically comprises the following steps:
Step one: the MRI image slice browsing the N number of patient suffering from brain tumor, one deck that its midbrain tumors is maximum is chosen from each patient, carry out the pre-service such as denoising, enhancing to it, and brain tumor split, adjustment obtains the gray level image of one group of N number of brain tumor;
Step 2: the textural characteristics extracting brain tumor based on Gabor wavelet, gets scale parameter V, direction number U during extraction, utilize wavelet basis to extract the Gabor wavelet texture feature vector g of V × U dimension to the gray level image of a brain tumor
1,1, g
1,2..., g
v,U, they are coupled together and obtain Gabor characteristic vector:
Calculate each tumour and obtain a stack features vector G
1, G
2..., G
n;
Step 3: to the gray level image of each tumour, the center of getting it is as the center of circle, and for each tumour constructs one group of k radius and radius increases the concentric circles of d at every turn, namely each radius of a circle is followed successively by d from inside to outside, 2d, 3d ..., kd.To each radius increment, calculate the length that tumour covers this rounded edge, and calculate the ratio of this length and this circumference, form the k dimensional feature vector of each tumour, obtain a stack features vector H of all tumours
1, H
2..., H
n;
Step 4: (1) setting entire deep learning model is the degree of deep learning network comprising L hidden layer, by the textural characteristics G in step 2
1, G
2..., G
ncombination obtains input vector X:
(2) utilize the degree of depth to learn, first carry out pre-training successively, use x as input vector, first utilize a parameter for θ={ determination of W, b} maps y=f
θx input x is mapped as a hiding potential expression y by ()=s (Wx+b), wherein W is weight matrix, and b is offset vector;
(3) the potential expression y obtained is θ '={ W ', b ' } function z=g by a parameter again
θ 'y ()=s (W ' y+b ') maps back the reconstruction vector z, the W ' meet constraint condition W '=W that become in an input space
t;
(4) each training sample x
iall be mapped to as corresponding y
i, reconstruct becomes z again
i, then by optimized reconstruction model parameter, average reconstruction error is minimized:
Wherein L
hbe loss function, rebuild cross entropy, that is:
(5) stochastic gradient descent is utilized to adjust middle parameter, error after making original input and rebuilding is minimum, obtain best parameter, the output of first hidden layer is obtained by this parameter, as the input of second hidden layer, by that analogy, until parameter after obtaining L hidden layer degree of deep learning process;
Step 5: as Fig. 3, finally increases an output layer at the network of pre-training, by there being the data of label, utilizing backpropagation and gradient descent method to finely tune the parameter of whole network from top to bottom, obtaining final argument;
Step 6: with the output G ' of network as the high-level characteristic extracted, augmented feature vector J=(G, G ', H) is connected to form with low layer textural characteristics G and shape facility H, with Data Dimensionality Reduction Algorithm PCA, dimensionality reduction calculating is carried out to this proper vector, obtain the final feature of all samples;
Step 7: by all proper vector J after dimensionality reduction
i, i=1,2 ..., N inputs support vector machine, training classifier;
Step 8: extract same feature according to feature extracting method identical before to test data, obtains the proper vector of all test datas, is entered in sorter, obtains classification results, statistics accuracy rate.
Above the brain tumor sorting technique based on degree of depth study that the present invention proposes is described in detail, but the description of embodiment is only for explaining method of the present invention and core concept thereof, so that the technician of this technology neck understands the present invention, but should be clear, the invention is not restricted to the scope of embodiment, to those skilled in the art, as long as various change to limit and in the spirit and scope of the present invention determined in appended claim, these changes are apparent, all innovation and creation utilizing the present invention to conceive are all at the row of protection.
Claims (3)
1., based on a brain tumor sorting technique for degree of depth study, it is characterized in that, comprise the following steps:
Step one: the magnetic resonance imaging (Magnetic ResonanceImaging, MRI) browsing the patient suffering from brain tumor, chooses one deck that its midbrain tumors is maximum, carries out the pre-service such as denoising, enhancing, and brain tumor split it;
Step 2: the textural characteristics extracting brain tumor based on Gabor wavelet, extracts the Gabor wavelet textural characteristics on different scale and direction;
Step 3: for each tumour constructs the concentric circles of one group of k radius equivalent increase, to each radius increment, calculate the ratio of this circle and tumour lap, form the k dimensional feature vector of each tumour;
Step 4: setting entire deep learning model is the degree of deep learning network comprising L hidden layer, using the texture feature vector in step 2 as input, the degree of depth is utilized to learn to carry out the extraction of high-level characteristic, first pre-training is successively carried out, utilize noise reduction automatic coding, coding and decoding is carried out to input, adjustment parameter between layers, error after making original input and rebuilding is minimum, the output of first hidden layer is obtained by this parameter, as the input of second hidden layer, by that analogy, until parameter after obtaining L hidden layer degree of deep learning process;
Step 5: finally increase an output layer at the network of pre-training, by there being the data of label, utilizing backpropagation and gradient descent method to finely tune the parameter of whole network from top to bottom, obtaining final argument;
Step 6: with the output of degree of deep learning network as the high-level characteristic extracted, be combined with low layer textural characteristics and shape facility again and form augmented feature vector, analysis (Principal ComponentAnalysis is become into main, PCA) dimensionality reduction calculating is carried out to this proper vector, thus obtain the final feature of all samples;
Step 7: using the proper vector after dimensionality reduction as input, puts into support vector machine training classifier;
Step 8: extract same feature according to feature extracting method identical before to test data, obtains the proper vector of all test datas, is entered in sorter, obtains classification results, statistics accuracy rate.
2. the brain tumor sorting technique based on degree of depth study according to claim 1, is characterized in that: described in described step 4 carries out high-level characteristic extraction to textural characteristics, and concrete steps are as follows:
(1) setting entire deep learning model is the degree of deep learning network comprising L hidden layer, by the textural characteristics G in step 2
1, G
2..., G
ncombination obtains input vector X:
(2) utilize the degree of depth to learn, first carry out pre-training successively, use x as input vector, first utilize a parameter for θ={ determination of W, b} maps y=f
θx input x is mapped as a hiding potential expression y by ()=s (Wx+b), wherein W is weight matrix, and b is offset vector;
(3) the potential expression y obtained is θ '={ W ', b ' } function z=g by a parameter again
θ 'y ()=s (W ' y+b ') maps back reconstruction vector z, W ' meet constraint condition W '=W in an input space
t;
(4) each training sample x
iall be mapped to as corresponding y
i, reconstruct becomes z again
i, then by optimized reconstruction model parameter, average reconstruction error is minimized:
Wherein L
hbe loss function, rebuild cross entropy, that is:
(5) stochastic gradient descent is utilized to adjust middle parameter, error after making original input and rebuilding is minimum, obtain best parameter, the output of first hidden layer is obtained by this parameter, as the input of second hidden layer, by that analogy, until parameter after obtaining L hidden layer degree of deep learning process.
3. as claimed in claim 1 a kind of based on the degree of depth study brain tumor sorting technique, it is characterized in that: the brain tumor classification concrete steps in described step 8 are, brain tumor image is input to brain tumor degree of depth network, obtain the high-level characteristic of brain tumor, to be combined with shape facility with textural characteristics and to form last proper vector, input and obtain classification results into support vector machine classifier.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510268230.0A CN104834943A (en) | 2015-05-25 | 2015-05-25 | Brain tumor classification method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510268230.0A CN104834943A (en) | 2015-05-25 | 2015-05-25 | Brain tumor classification method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104834943A true CN104834943A (en) | 2015-08-12 |
Family
ID=53812820
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510268230.0A Pending CN104834943A (en) | 2015-05-25 | 2015-05-25 | Brain tumor classification method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104834943A (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106023188A (en) * | 2016-05-17 | 2016-10-12 | 天津大学 | Breast tumor feature selection method based on Relief algorithm |
| CN106127747A (en) * | 2016-06-17 | 2016-11-16 | 史方 | Car surface damage classifying method and device based on degree of depth study |
| CN106156530A (en) * | 2016-08-03 | 2016-11-23 | 北京好运到信息科技有限公司 | Health check-up data analysing method based on stack own coding device and device |
| CN106204532A (en) * | 2016-06-24 | 2016-12-07 | 华南理工大学 | The sorting technique of the tumor of feature based data mining and neutral net |
| CN106504253A (en) * | 2016-10-14 | 2017-03-15 | 国政通科技股份有限公司 | A kind of processing method of medical imaging photo and system |
| CN106682435A (en) * | 2016-12-31 | 2017-05-17 | 西安百利信息科技有限公司 | System and method for automatically detecting lesions in medical image through multi-model fusion |
| CN106780453A (en) * | 2016-12-07 | 2017-05-31 | 电子科技大学 | A kind of method realized based on depth trust network to brain tumor segmentation |
| CN106780499A (en) * | 2016-12-07 | 2017-05-31 | 电子科技大学 | A kind of multi-modal brain tumor image partition method based on stacking autocoding network |
| CN106971198A (en) * | 2017-03-03 | 2017-07-21 | 北京市计算中心 | A kind of pneumoconiosis grade decision method and system based on deep learning |
| CN106971410A (en) * | 2017-03-27 | 2017-07-21 | 华南理工大学 | A kind of white matter fiber tract method for reconstructing based on deep learning |
| CN107220980A (en) * | 2017-05-25 | 2017-09-29 | 重庆理工大学 | A kind of MRI image brain tumor automatic division method based on full convolutional network |
| CN107260159A (en) * | 2017-06-28 | 2017-10-20 | 浙江理工大学 | A kind of cardiac electric functional imaging method based on convolutional neural networks |
| CN107506796A (en) * | 2017-08-25 | 2017-12-22 | 电子科技大学 | A kind of alzheimer disease sorting technique based on depth forest |
| CN108416360A (en) * | 2018-01-16 | 2018-08-17 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification feature |
| CN108898175A (en) * | 2018-06-26 | 2018-11-27 | 北京工业大学 | Area of computer aided model building method based on deep learning gastric cancer pathological section |
| CN109558774A (en) * | 2017-09-27 | 2019-04-02 | 中国海洋大学 | Object automatic recognition system based on depth residual error network and support vector machines |
| CN109919187A (en) * | 2019-01-28 | 2019-06-21 | 浙江工商大学 | With bagging fine tuning CNN come the method for Thyroid Follicular picture of classifying |
| CN110188813A (en) * | 2019-05-24 | 2019-08-30 | 上海联影智能医疗科技有限公司 | Characteristics of image classification method, computer equipment and storage medium |
| WO2019170711A1 (en) | 2018-03-07 | 2019-09-12 | Institut National De La Sante Et De La Recherche Medicale (Inserm) | Method for early prediction of neurodegenerative decline |
| CN111340767A (en) * | 2020-02-21 | 2020-06-26 | 四川大学华西医院 | Method and system for processing scalp positioning image of brain tumor |
| CN111353998A (en) * | 2020-05-13 | 2020-06-30 | 温州医科大学附属第一医院 | Tumor diagnosis and treatment prediction model and device based on artificial intelligence |
| CN111832563A (en) * | 2020-07-17 | 2020-10-27 | 江苏大学附属医院 | Intelligent breast tumor identification method based on ultrasonic image |
| US10939874B2 (en) | 2016-03-24 | 2021-03-09 | The Regents Of The University Of California | Deep-learning-based cancer classification using a hierarchical classification framework |
| CN117437493A (en) * | 2023-12-20 | 2024-01-23 | 泰山学院 | Brain tumor MRI image classification method and system combining first-order and second-order features |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006114003A1 (en) * | 2005-04-27 | 2006-11-02 | The Governors Of The University Of Alberta | A method and system for automatic detection and segmentation of tumors and associated edema (swelling) in magnetic resonance (mri) images |
| CN102201038A (en) * | 2011-04-27 | 2011-09-28 | 浙江大学 | Method for detecting P53 protein expression in brain tumor |
| CN102596025A (en) * | 2010-06-25 | 2012-07-18 | 中国科学院自动化研究所 | Method and system for brain tumor segmentation in multi-parameter 3d mr images via robust statistic information propagation |
| US8280133B2 (en) * | 2008-08-01 | 2012-10-02 | Siemens Aktiengesellschaft | Method and system for brain tumor segmentation in 3D magnetic resonance images |
| CN103366180A (en) * | 2013-06-14 | 2013-10-23 | 山东大学 | Cell image segmentation method based on automatic feature learning |
-
2015
- 2015-05-25 CN CN201510268230.0A patent/CN104834943A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2006114003A1 (en) * | 2005-04-27 | 2006-11-02 | The Governors Of The University Of Alberta | A method and system for automatic detection and segmentation of tumors and associated edema (swelling) in magnetic resonance (mri) images |
| US8280133B2 (en) * | 2008-08-01 | 2012-10-02 | Siemens Aktiengesellschaft | Method and system for brain tumor segmentation in 3D magnetic resonance images |
| CN102596025A (en) * | 2010-06-25 | 2012-07-18 | 中国科学院自动化研究所 | Method and system for brain tumor segmentation in multi-parameter 3d mr images via robust statistic information propagation |
| CN102201038A (en) * | 2011-04-27 | 2011-09-28 | 浙江大学 | Method for detecting P53 protein expression in brain tumor |
| CN103366180A (en) * | 2013-06-14 | 2013-10-23 | 山东大学 | Cell image segmentation method based on automatic feature learning |
Non-Patent Citations (1)
| Title |
|---|
| 雷成 等: "深度学习技术及其在肿瘤分类中的应用", 《智能计算机与应用》 * |
Cited By (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10939874B2 (en) | 2016-03-24 | 2021-03-09 | The Regents Of The University Of California | Deep-learning-based cancer classification using a hierarchical classification framework |
| CN106023188A (en) * | 2016-05-17 | 2016-10-12 | 天津大学 | Breast tumor feature selection method based on Relief algorithm |
| CN106127747B (en) * | 2016-06-17 | 2018-10-16 | 史方 | Car surface damage classifying method and device based on deep learning |
| CN106127747A (en) * | 2016-06-17 | 2016-11-16 | 史方 | Car surface damage classifying method and device based on degree of depth study |
| CN106204532A (en) * | 2016-06-24 | 2016-12-07 | 华南理工大学 | The sorting technique of the tumor of feature based data mining and neutral net |
| CN106204532B (en) * | 2016-06-24 | 2019-01-18 | 华南理工大学 | The classification method of tumour based on characteristic excavation and neural network |
| CN106156530A (en) * | 2016-08-03 | 2016-11-23 | 北京好运到信息科技有限公司 | Health check-up data analysing method based on stack own coding device and device |
| CN106504253A (en) * | 2016-10-14 | 2017-03-15 | 国政通科技股份有限公司 | A kind of processing method of medical imaging photo and system |
| CN106780453A (en) * | 2016-12-07 | 2017-05-31 | 电子科技大学 | A kind of method realized based on depth trust network to brain tumor segmentation |
| CN106780499A (en) * | 2016-12-07 | 2017-05-31 | 电子科技大学 | A kind of multi-modal brain tumor image partition method based on stacking autocoding network |
| CN106682435A (en) * | 2016-12-31 | 2017-05-17 | 西安百利信息科技有限公司 | System and method for automatically detecting lesions in medical image through multi-model fusion |
| CN106682435B (en) * | 2016-12-31 | 2021-01-29 | 西安百利信息科技有限公司 | System and method for automatically detecting lesion in medical image through multi-model fusion |
| CN106971198A (en) * | 2017-03-03 | 2017-07-21 | 北京市计算中心 | A kind of pneumoconiosis grade decision method and system based on deep learning |
| CN106971410A (en) * | 2017-03-27 | 2017-07-21 | 华南理工大学 | A kind of white matter fiber tract method for reconstructing based on deep learning |
| CN107220980B (en) * | 2017-05-25 | 2019-12-03 | 重庆师范大学 | A kind of MRI image brain tumor automatic division method based on full convolutional network |
| CN107220980A (en) * | 2017-05-25 | 2017-09-29 | 重庆理工大学 | A kind of MRI image brain tumor automatic division method based on full convolutional network |
| CN107260159A (en) * | 2017-06-28 | 2017-10-20 | 浙江理工大学 | A kind of cardiac electric functional imaging method based on convolutional neural networks |
| CN107506796A (en) * | 2017-08-25 | 2017-12-22 | 电子科技大学 | A kind of alzheimer disease sorting technique based on depth forest |
| CN109558774A (en) * | 2017-09-27 | 2019-04-02 | 中国海洋大学 | Object automatic recognition system based on depth residual error network and support vector machines |
| CN108416360A (en) * | 2018-01-16 | 2018-08-17 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification feature |
| CN108416360B (en) * | 2018-01-16 | 2022-03-29 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification features |
| WO2019170711A1 (en) | 2018-03-07 | 2019-09-12 | Institut National De La Sante Et De La Recherche Medicale (Inserm) | Method for early prediction of neurodegenerative decline |
| CN108898175B (en) * | 2018-06-26 | 2020-11-20 | 北京工业大学 | Computer-aided model construction method based on deep learning gastric cancer pathological section |
| CN108898175A (en) * | 2018-06-26 | 2018-11-27 | 北京工业大学 | Area of computer aided model building method based on deep learning gastric cancer pathological section |
| CN109919187A (en) * | 2019-01-28 | 2019-06-21 | 浙江工商大学 | With bagging fine tuning CNN come the method for Thyroid Follicular picture of classifying |
| CN109919187B (en) * | 2019-01-28 | 2021-02-12 | 浙江工商大学 | Method for classifying thyroid follicular picture by using bagging fine tuning CNN |
| CN110188813A (en) * | 2019-05-24 | 2019-08-30 | 上海联影智能医疗科技有限公司 | Characteristics of image classification method, computer equipment and storage medium |
| CN111340767A (en) * | 2020-02-21 | 2020-06-26 | 四川大学华西医院 | Method and system for processing scalp positioning image of brain tumor |
| CN111340767B (en) * | 2020-02-21 | 2023-12-12 | 四川大学华西医院 | Brain tumor scalp positioning image processing method and system |
| CN111353998A (en) * | 2020-05-13 | 2020-06-30 | 温州医科大学附属第一医院 | Tumor diagnosis and treatment prediction model and device based on artificial intelligence |
| CN111832563A (en) * | 2020-07-17 | 2020-10-27 | 江苏大学附属医院 | Intelligent breast tumor identification method based on ultrasonic image |
| CN117437493A (en) * | 2023-12-20 | 2024-01-23 | 泰山学院 | Brain tumor MRI image classification method and system combining first-order and second-order features |
| CN117437493B (en) * | 2023-12-20 | 2024-03-29 | 泰山学院 | Brain tumor MRI image classification method and system combining first-order and second-order features |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104834943A (en) | Brain tumor classification method based on deep learning | |
| Elguindi et al. | Deep learning-based auto-segmentation of targets and organs-at-risk for magnetic resonance imaging only planning of prostate radiotherapy | |
| Chen et al. | Fully automatic acute ischemic lesion segmentation in DWI using convolutional neural networks | |
| Işın et al. | Review of MRI-based brain tumor image segmentation using deep learning methods | |
| Men et al. | Automatic segmentation of the clinical target volume and organs at risk in the planning CT for rectal cancer using deep dilated convolutional neural networks | |
| Mansoor et al. | Deep learning guided partitioned shape model for anterior visual pathway segmentation | |
| Pang et al. | A deep model for lung cancer type identification by densely connected convolutional networks and adaptive boosting | |
| CN109523521A (en) | Lung neoplasm classification and lesion localization method and system based on more slice CT images | |
| CN107220966A (en) | A kind of Histopathologic Grade of Cerebral Gliomas Forecasting Methodology based on image group | |
| Khan et al. | Recent automatic segmentation algorithms of MRI prostate regions: a review | |
| Piantadosi et al. | DCE-MRI breast lesions segmentation with a 3TP U-Net deep convolutional neural network | |
| CN106408001A (en) | Rapid area-of-interest detection method based on depth kernelized hashing | |
| Xiao et al. | Glioma grading prediction by exploring radiomics and deep learning features | |
| Reda et al. | A new NMF-autoencoder based CAD system for early diagnosis of prostate cancer | |
| Jia et al. | Prostate segmentation in MR images using ensemble deep convolutional neural networks | |
| Alkassar et al. | Automatic brain tumour segmentation using fully convolution network and transfer learning | |
| Zhang et al. | Concatenated spatially-localized random forests for hippocampus labeling in adult and infant MR brain images | |
| CN106651874A (en) | Brain tumor post-operation spatial domain partitioning method based on multimode MRI data | |
| Xu et al. | Mammographic mass segmentation using multichannel and multiscale fully convolutional networks | |
| Mushtaq et al. | Brain Tumor Segmentation using Multi-View Attention based Ensemble Network. | |
| Lei et al. | Deep learning architecture design for multi-organ segmentation | |
| Micallef et al. | A nested U-net approach for brain tumour segmentation | |
| Jiang et al. | Deep cross‐modality (MR‐CT) educed distillation learning for cone beam CT lung tumor segmentation | |
| Qian et al. | ProCDet: A new method for prostate cancer detection based on mr images | |
| Rosas González et al. | 3D automatic brain tumor segmentation using a multiscale input U-Net network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20150812 |
|
| WD01 | Invention patent application deemed withdrawn after publication |