Accurate classification on pathological images is a significant research focus such as for non-Hodgkin lymphomas (NHL). To this end, this paper proposes a hierarchical classification model based on the labels' sta...
详细信息
ISBN:
(纸本)9781538666500
Accurate classification on pathological images is a significant research focus such as for non-Hodgkin lymphomas (NHL). To this end, this paper proposes a hierarchical classification model based on the labels' statistics for three NHL pathological images, including chronic lymphocytic leukemia (CLL), follicular lymphoma (FL) and mantle cell lymphoma (MCL). First, each pathological image is converted onto the grayscale channel and then divided into 130 non-overlapped patches with 100 x 100 pixels. Next, the sparse autoencoder (SAE), an unsupervised feature extraction method, is utilized to learn the representations of all patches and meanwhile texture features are extracted on these patches which are considered as the hand-craft features. Following this process, we can obtain a 680-dimension feature set. Finally, a hierarchical classification model trained by these 680-dimension features is applied to classify NHL as CLL, FL and MCL, where the label of each NHL pathological image is determined via the output labels of its 130 patches. The experimental results and comparisons demonstrate the advantages of the proposed hierarchical classification model.
When a fracturing vehicle is working, it generally needs to bear high loads, media corrosion and erosion. For this special working environment, this study proposes a rolling bearing fault diagnosis method based on sta...
详细信息
When a fracturing vehicle is working, it generally needs to bear high loads, media corrosion and erosion. For this special working environment, this study proposes a rolling bearing fault diagnosis method based on stack marginalised sparse denoising auto-encoder (SDAE). This method combines the sparse auto-encoder (SAE) and the denoising auto-encoder (DAE) and combines the characteristics of dimensionality reduction and robustness. The method adds marginalisation to optimise the SDAE. Finally, it uses a two-layer stacking method. The output results of the second marginalised SDAE are used as input to the softmax classifier for learning training and classification testing. This improved method (stack SDAE) improves the denoising ability, reduces the computational complexity, solves the problems of difficult parameter adjustment and slows training convergence. The experimental tests were carried out on the failure of pitting corrosion of the outer ring of the bearing, pitting failure of the inner ring, and cracking of the rolling element. The results show that the algorithm can effectively improve the accuracy of fault diagnosis of rolling bearings, and it has greatly improved than the algorithms of SAEs and DAE.
In recent years, deep learning has been extensively used in both supervised and unsupervised learning problems. Among the deep learning models, CNN has outperformed all others for object recognition task. Although CNN...
详细信息
In recent years, deep learning has been extensively used in both supervised and unsupervised learning problems. Among the deep learning models, CNN has outperformed all others for object recognition task. Although CNN achieves exceptional accuracy, still a huge number of iterations and chances of getting stuck in local optima makes it computationally expensive to train. Genetic Algorithm is a metaheuristic approach inspired by the theory of natural selection and has been used for solving both bounded and unbounded optimization problems by a large success. To handle these issues, we have developed a hybrid deep learning model using Genetic Algorithm and L-BFGS method for training CNN. To test our model, we have taken the Devanagari handwritten numeral dataset. Our results show that GA assisted CNN produces better results than non-GA assisted CNN. This study concludes that evolutionary technique can be used to train CNN more efficiently. (C) 2018 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the scientific committee of the 6th International Conference on Smart Computing and Communications.
Pan-sharpening is a common image-fusion method. To improve the quality of fused images, a multilevel deep learning Pan-sharpening method is proposed in this paper. In the training phase, we introduce Coupled sparse De...
详细信息
ISBN:
(纸本)9781538671504
Pan-sharpening is a common image-fusion method. To improve the quality of fused images, a multilevel deep learning Pan-sharpening method is proposed in this paper. In the training phase, we introduce Coupled sparse Denoising Autoencorder (CSDA) to reconstruct high-Resolution (HR) multispectral (MS) image from low-Resolution (LR) MS image and HR Panchromatic (Pan) image. CSDA has four networks including LM-HP network, HR-MS network, feature mapping network and fine-tuning network. The hidden features in LM-HP network and HR-MS network as well as the mapping function between the two features are learned through joint optimization. In LM-HP and HR-MS networks, the hidden features of image patch pairs are extracted by the sparse autoencoder. A sparse denoising autoencoder is used to build the nonlinear mapping between the extracted features. In the testing phase, the LR-MS and HR-Pan images patches are fed to the CSDA network to reconstruct the fused HR-MS image. The experimental results show that the proposed method is better than the traditional pans-sharpening methods.
sparse autoencoder is a commonly used deep learning approach for automatically learning features from unlabelled data (unsupervised feature learning). This paper proposes class-specific (supervised) pre-trained approa...
详细信息
ISBN:
(纸本)9781509020508
sparse autoencoder is a commonly used deep learning approach for automatically learning features from unlabelled data (unsupervised feature learning). This paper proposes class-specific (supervised) pre-trained approach based on sparse autoencoder to gain low-dimensional interesting structure of features with high performance in document classification. Experimental results have demonstrated the advantages and usefulness of the proposed method in document classification in high-dimensional feature space, in terms of the limited number of features required to achieve good classification accuracy.
To effectively improve the diagnosis of pulmonary nodules, this paper proposes a new automatic diagnosis method for pulmonary nodules based on a new hierarchical extreme learning machine (H-ELM) that can automatically...
详细信息
To effectively improve the diagnosis of pulmonary nodules, this paper proposes a new automatic diagnosis method for pulmonary nodules based on a new hierarchical extreme learning machine (H-ELM) that can automatically carry out feature extraction, model training and pulmonary nodule detection. In our method, an adaptive histogram equalisation is used first to enhance contrast of the original pulmonary nodule image. The processed images are then input into an extreme learning machine (ELM)-based unsupervised multilayer auto-encoder to obtain more compact and meaningful high-level features of the pulmonary nodule image. Finally, supervised feature classification, which uses these high-level features of the pulmonary nodule as input data, is implemented using the ELM classifier. In the experiments, 2,800 pulmonary nodule images are used to validate the proposed method, and compared with existing pulmonary nodule diagnosis methods, our proposed method is more accurate and less time consuming and effectively avoids the complexity of manual feature extraction.
With millions of people suffering from dementia worldwide, the global prevalence of dementia has a significant impact on the patients' lives, their caregivers' physical and emotional states, and the global eco...
详细信息
ISBN:
(纸本)9781538616451
With millions of people suffering from dementia worldwide, the global prevalence of dementia has a significant impact on the patients' lives, their caregivers' physical and emotional states, and the global economy. Early diagnosis of dementia helps in finding suitable therapies that reduce or even prevent further deterioration of patients' cognitive abilities. MRI scans are shown to be the most effective strategy to enable early diagnosis of dementia. Nevertheless, the early diagnosis of dementia is a challenging task due to the high dimensionality of MRI scans that may degrade the effectiveness of machine learning models. Feature abstraction is a part of dimensionality reduction process need by researchers to represent input data in its simplest form that results in a more-robust system model. It is a technique that collects relevant features and ignores irrelevant or redundant ones from data without loss of much key information. This paper proposes a novel feature abstraction method using sparse autoencoder (SAE) to reduce the dimensionality of and extract key features from MRI neuroimages. These features and that obtained from the popular PCA approach are then used to train a Linear Discriminant Analysis (LDA) and Logistic Regression classifiers in order to compare their prediction accuracy. The experimental results show that the proposed approach yields higher classification accuracy compared to that using the PCA by 8 percent of classification accuracy. The experimental results show that the use of features learned by SAE in early diagnosis of multi-type dementia provides better classification performance than the use of raw image pixel intensities for diagnosing dementia.
This paper describes an artificial neural network (ANN) method that employs a feature-learning algorithm to detect the lumen and MA borders in intravascular ultrasound (IVUS) images. Three types of imaging features in...
详细信息
ISBN:
(纸本)9781509011728
This paper describes an artificial neural network (ANN) method that employs a feature-learning algorithm to detect the lumen and MA borders in intravascular ultrasound (IVUS) images. Three types of imaging features including spatial, neighboring, and gradient features were used as the input features to the neural network, and then the different vascular layers were distinguished using two sparse autoencoders and one softmax classifier. To smooth the lumen and MA borders detected by the ANN method, we used the active contour model. The performance of our approach was compared with the manual drawing method and another existing method on 538 IVUS images from six subjects. Results showed that our approach had a high correlation (r = 0.9284 +/- 0.9875 for all measurements) and good agreement (bias = 0.0148 +/- 0.4209 mm) with the manual drawing method, and small detection error (lumen border: 0.0928 +/- 0.0935 mm, MA border: 0.1056 +/- 0.1088 mm). The average time to process each image was 14 +/- 4.6 seconds. The obtained results indicate that our proposed approach can be used to efficiently and accurately detect the lumen and MA borders in IVUS images.
The identification of a cover song, which is an alternative version of a previously recorded song, for music retrieval has received increasing attention. Methods for identifying a cover song typically involve comparin...
详细信息
The identification of a cover song, which is an alternative version of a previously recorded song, for music retrieval has received increasing attention. Methods for identifying a cover song typically involve comparing the similarity of chroma features between a query song and another song in the data set. However, considerable time is required for pairwise comparisons. In this study, chroma features were patched to preserve the melody. An intermediate representation was trained to reduce the dimension of each patch of chroma features. The training was performed using an autoencoder, commonly used in deep learning for dimensionality reduction. Experimental results showed that the proposed method achieved better accuracy for identification and spent less time for similarity matching in both covers80 dataset and Million Song Dataset as compared with traditional approaches.
作者:
Gong, MaoguoYang, HailunZhang, PuzhaoXidian Univ
Minist Educ Int Res Ctr Intelligent Percept & Computat Key Lab Intelligent Percept & Image Understanding Xian 710071 Shaanxi Provinc Peoples R China
Ternary change detection aims to detect changes and group the changes into positive change and negative change. It is of great significance in the joint interpretation of spatial-temporal synthetic aperture radar imag...
详细信息
Ternary change detection aims to detect changes and group the changes into positive change and negative change. It is of great significance in the joint interpretation of spatial-temporal synthetic aperture radar images. In this study, sparse autoencoder, convolutional neural networks (CNN) and unsupervised clustering are combined to solve ternary change detection problem without any supervison. Firstly, sparse autoencoder is used to transform log-ratio difference image into a suitable feature space for extracting key changes and suppressing outliers and noise. And then the learned features are clustered into three classes, which are taken as the pseudo labels for training a CNN model as change feature classifier. The reliable training samples for CNN are selected from the feature maps learned by sparse autoencoder with certain selection rules. Having training samples and the corresponding pseudo labels, the CNN model can be trained by using back propagation with stochastic gradient descent. During its training procedure, CNN is driven to learn the concept of change, and more powerful model is established to distinguish different types of changes. Unlike the traditional methods, the proposed framework integrates the merits of sparse autoencoder and CNN to learn more robust difference representations and the concept of change for ternary change detection. Experimental results on real datasets validate the effectiveness and superiority of the proposed framework. (C) 2017 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS). Published by Elsevier B.V. All rights reserved.
暂无评论