Применение искусственных нейронных сетей для классификации изображений глаз с катарактой, полученных с помощью фотощелевой лампы

Multimodal medical information fusion has emerged as a revolutionary method in intelligent healthcare, allowing complete consideration of patient well-being and tailored treatment strategies. On the other hand, the current approach produces erroneous findings and has problems with the early phases of brain tumour prediction in MRI images. In healthcare, accurate and reliable brain classification of images is essential for diagnosis and strategic decision-making. Currently, semantic gaps are the main problem with brain tumour image classification. To fill the research gap, traditional ML models for classification use handcrafted features, which are low-level yet relatively high-level, and they use intensive approaches for feature extraction and classification. In recent years, substantial improvements have been made in deep learning for automated image classification. Recurrent Neural Networks (RNNs), or deep Convolutional Neural Networks (CNN), have been particularly effective in this multimodal image classification. Hence, this paper presents the Multimodal Fusion Model-assisted Convolutional Neural Network and Recurrent Neural Networks (MFM-CNN-RNN) for automatic image classification in smart healthcare. This study aims to determine if a fusion of CT and MRI brain scans is normal or abnormal. To enhance the accuracy of brain tumour image classification, this method uses the multimodality information within CNNs and RNNs by extracting and fusing unique and complimentary features from different modalities. Within this framework, features have been retrieved using CNN features, while dependencies and classification have been determined using RNN attributes. Because of its design, LSTM excels in time series analysis, which involves processing data in sequential order.

Iterative fusion convolutional neural networks for classification of optical coherence tomography images

Backgraund. Optical coherence tomography (OCT) is a modern high-tech and informative method for detecting pathology of the retina and preretinal layers of the vitreous body. However, the description and interpretation of the research results require high qualifications and special training of an ophthalmologist, and significant time expenditure for the doctor and the patient. At the same time, the use of mathematical models based on artificial neural networks (ANN- models) currently makes it possible to automate many processes associated with image processing. Therefore, solving problems associated with automating the process of classifying OCT images based on ANN models is actual.
 Aims. To develop architectures of mathematical (computer) models based on deep learning of convolutional neural networks (CNN) for classification of OCT images of the retina. To compare the results of computational experiments conducted using Python tools in the Google Colaboratory with single-model and multi-model approaches and evaluate classification accuracy. To make conclusions about the optimal architecture of ANN models and the values of the hyperparameters used.
 Materials and methods. The original dataset, which was anonymized OCT images of real patients, included more than 2000 images obtained directly from the device in a resolution of 1920 × 969 × 24 BPP. The number of image classes is 12. To create the training and validation data sets, a subject area of 1100 × 550 × 24 BPP was “cut out.” Various approaches were studied: the possibility of using pretrained CNNs with transfer learning, techniques for resizing and augmenting images, as well as various combinations of hyperparameters of ANN-models. When compiling the model, the following parameters were used: Adam optimizer, categorical_crossentropy loss function, accuracy metric. All technological processes with images and ANN-models were carried out using Python language tools in Google Colaboratory.
 Results. Single-model and multi-model principles for classifying OCT images of the retina are proposed. Computational experiments on automated classification of such images obtained from a DRI OCT Triton 3D tomograph using various ANN model architectures showed an accuracy of 98-100% during training and validation and 85% during an additional test, which is a satisfactory result. The optimal architecture of the ANN model - a six-layer convolutional network - was selected and the values of its hyperparameters were determined.
 Conclusions. The results of deep training of convolutional neural network models with various architectures, their validation and testing showed satisfactory classification accuracy of retinal OCT images. These developments can be used in decision support systems in the field of ophthalmology.

Ultrasound imaging has become a routine means of diagnosing atherosclerosis. The classification of carotid ultrasound images and detection for the plaques automatically are critical for the diagnosis of atherosclerosis, which has important clinical significance for further analysis of plaque vulnerability and risk assessment of cardiovascular and cerebrovascular events. At present, manual measurement is used for the classification, which has obvious disadvantages such as inaccurate measurement and operator variability. In this paper, we proposed an automatic classification method based on convolutional neural network (CNN) for the carotid ultrasound images from different research institutions and ultrasound machines. 820 and 830 carotid ultrasound images from Zhongnan Hospital of Wuhan University and Robarts Research Institute of Canada were used for the classification of normal, thickened vessel wall and plaque images. To solve the problem of uneven image quality and size, we used six different image normalization schemes. Furthermore, we designed five CNNs with slightly different structures and compared them with texture-based features classifications. The CNN results showed significant superiority in classification performance with total accuracy of 90.30% and recall rate of 89.70%, indicating the automatic classification of carotid ultrasound images based on CNN is potentially useful for clinical application in the diagnosis of carotid atherosclerosis.

The impressive gain in performance obtained using deep neural networks (DNN) for various tasks encouraged us to apply DNN for image classification task. We have used a variant of DNN called Deep convolutional Neural Networks (DCNN) for feature extraction and image classification. Neural networks can be used for classification as well as for feature extraction. Our whole work can be better seen as two different tasks. In the first task, DCNN is used for feature extraction and classification task. In the second task, features are extracted using DCNN and then SVM, a shallow classifier, is used to classify the extracted features. Performance of these tasks is compared. Various configurations ofDCNNare used for our experimental studies.Among different architectures that we have considered, the architecture with 3 levels of convolutional and pooling layers, followed by a fully connected output layer is used for feature extraction. In task 1 DCNN extracted features are fed to a 2 hidden layer neural network for classification. In task 2 SVM is used to classify the features extracted by DCNN. Experimental studies show that the performance of υ-SVM classification on DCNN features is slightly better than the results of neural network classification on DCNN extracted features.

Abstract. This paper reviews the application and improvement of convolutional neural networks (CNNs) in image classification. Firstly, a shallow CNN for interstitial lung disease image classification is presented. This model suppresses overfitting through a unique network architecture and optimisation algorithm. Next, the improved VGG16 architecture and MIDNet18 model are discussed and their superior performance in brain tumour image classification is demonstrated. Subsequently, a CNN-CapsNet model for cervical cancer image classification and its improvement are presented and the customised model is compared with the conventional VGG-16 CNN architecture in the paper. Next, the application of sparse convolutional kernels and hybrid sparse convolutional kernels (HDCs) in solving the problem of computational resource consumption is presented. Subsequently, methods for solving the problem of limited training data through transfer learning and network data augmentation techniques are discussed, as well as GAN-generated datasets for solving the overfitting problem. Finally, the effect of degraded images on the classification effectiveness of CNNs is explored. The results show that the improved CNN architecture and algorithms have significant effects in solving the problems of overfitting and computational resource consumption, and can significantly improve the accuracy and efficiency of image classification. And degraded images do adversely affect the accuracy of CNN for image classification.

This paper proposes a new model for automatic classification of breast cancer tissues images using convolution neural network on BreakHis dataset. The main characteristic of the proposed model is its independency on the magnification factors of the images. The presence of pooling layer only in the last convolutional layer is the beauty of this model, which assists in the prevention of information loss. Data augmentation technique was used to increase the size of the dataset as convolution neural network relies on the size of the dataset for its better performance. For model evaluation, the classification performance of the proposed model was compared with the recent work and found that the proposed model outperforms the existing one with an average accuracy of 85.3% as well as robust to the images with different magnification factor. Employment of additional data, deeper architecture and consideration of factors like filter size, pooling strategy, optimiser, loss function can be the future possibilities for this work.

Because the LeNet-5 convolutional neural network is not accurate in classifying images with complex texture features, in order to improve the accuracy, this paper proposes a multi-convolution neural network. First, on the basis of the LeNet-5 network, a convolutional layer and a pooling layer are added to deepen the depth of the network. Secondly, the first two pooling layers are propagated backward and connected to the last pooling layer through convolution, so as to make full use of the low-level features extracted by the network and effectively combining the low-level features and high-level features extracted by the network. Finally, use the Adam optimizer to adjust and update the parameters of the network. Experimental results on the Cats vs. Dogs, Cifar-10 and Fer2013 datasets show that the multi-convolution neural network has good classification ability on complex texture feature datasets.

A convolutional fuzzy min-max neural network

Convolutional neural network (CNN) has been successfully used in many fields including image recognition. CNN is composed of input, convolution, pooling, hidden and output layers, and the weights and biases between layers except the ones between convolution and pooling layers are acquired by learning. In comparison to the conventional neural networks, the learning cost of CNN is higher, and the learning time is longer especially when hidden layer(s) are added. Recently, complex- and quaternion-valued neural networks have drawn much attention. In complex-valued neural networks, inputs, weights, biases and outputs are complex numbers, and in quaternion-valued neural networks, these parameters are quaternions. It has been shown that both methods exhibit excellent accuracy in various applications such as classification and function approximation problems with less computational burden. In this study, we propose CNNs with complex- and quaternion-valued neurons where complex and quaternion numbers are used between pooling, hidden and output layers. We here show that CNNs with complex- and quaternion-valued neurons have higher learning ability in handwritten digit image classification with the MNIST data than the real-valued counterpart.

The application of deep learning technology has increased rapidly in recent years. Technologies in deep learning increasingly emulate natural human abilities, such as knowledge learning, problem-solving, and decision-making. In general, deep learning can carry out self-training without repetitive programming by humans. Convolutional neural networks (CNNs) are deep learning algorithms commonly used in wide applications. CNN is often used for image classification, segmentation, object detection, video processing, natural language processing, and speech recognition. CNN has four layers: convolution layer, pooling layer, fully connected layer, and non-linear layer. The convolutional layer uses kernel filters to calculate the convolution of the input image by extracting the fundamental features. The pooling layer combines two successive convolutional layers. The third layer is the fully connected layer, commonly called the convolutional output layer. The activation function defines the output of a neural network, such as 'yes' or 'no'. The most common and popular CNN activation functions are Sigmoid, Tanh, ReLU, Leaky ReLU, Noisy ReLU, and Parametric Linear Units. The organization and function of the visual cortex greatly influence CNN architecture because it is designed to resemble the neuronal connections in the human brain. Some of the popular CNN architectures are LeNet, AlexNet and VGGNet.

Deep feedforward network (DFN) is a conceptual stepping stone of many well-known deep neural networks (DNN) in image classification and natural language application. The development on the standard DFN can rarely be found in the literature recently due to the popularity in convolutional networks. The recent trend of research focuses on the increment of the convolutional layers in a deeper and wider network architecture for achieving higher accuracy and lower misclassification rate. However, stacking the convolutional layers may not result in better accuracy due to the sparsity of interconnected of hidden nodes. In this paper, a convolutional deep feedforward network (C-DFN) is proposed to anlayse the performance of deep neural networks by increasing the number of fully-connected layers. C-DFN contains a Gabor-convolutional layer as a trainable feature extractor and followed by the four fully-connected layers. Experiments are conducted to evaluate the performance of proposed network with three other structures, i.e. deep belief network, deep feedforward network and convolutional deep belief network. The experimental results showed that C-DFN obtained the lowest average misclassfication rate of 9.41% in the image classification.

With the continuous development of convolutional neural networks (CNNs), image classification technology has entered a new stage in solving visual cognition tasks. Recent advances have shown that if containing short connections, convolutional networks can be more accurate and efficient to train. However, these short connections almost only exist between convolutional layers, which potentially make the feature information flow insufficiently. In addition, the multi‐scale representation ability of CNNs using short connections can be further explored. Thus, the dense short connection network (DSCNet) for efficient image classification is designed in this paper. In DSCNet, we propose a simple yet effective architectural unit, namely dense short connection (DSC) module, which allows hierarchical dense short connections for multi‐scale context information across different feature channels within a single convolutional layer. DSCNets have several noteworthy advantages: they improve the multi‐scale representation ability, enhance feature propagation, and reduce the number of parameters. To validate our DSC, we conduct comprehensive experiments on CIFAR‐100 and Tiny ImageNet datasets. Experimental results show that DSCNets achieve very competitive results with previous baseline models, whilst requiring less computation to achieve higher recognition accuracy. Further ablation studies show that our proposed method can achieve consistent performance gains in image classification tasks.

The quality of classification is crucial in medical applications. Especially when it comes to confirm that the patient does not have a malignant tumor. An example of such an application is a binary classification of breast tumor malignancy based on histopathological images. This paper explains the most popular attention mechanism in convolution neural networks as follows. Convolutional Block Attention Module, Attention Augmented Convolution, and Attention Guided Convolutional Neural Networks. Four neural networks are built and compared. Each is evaluated in the classification problem of histopathological images of breast cancer. On the basis of the results, it is clear that some attentional neural networks can outperform standard convolutional networks in the classification of breast cancer. In our investigation, the convolution networks reached an accuracy level of 90% and an AUC-ROC of 95.9%. It is worse compared to the Convolutional Block Attention Module Network (accuracy 90.7%, AUC-ROC 96.9%) and the Attention-Guided Convolutional Network (accuracy 91.2%, AUC-ROC 96.6%). Attention-augmented convolution remains behind the standard convolutional network, achieving 88.9% accuracy and 94.8% AUC-ROC. The Attention-Guided Convolution Network was the best network of all four. We also compared precision, NPV, sensitivity, specificity, and \(F_{1}\)-score. We came to the conclusion that the Convolutional Block Attention Module network has the highest NPV (90.8%) and sensitivity (96.2%), while the Attention-Guided Convolutional Neural Network scored the highest in precision (92.4%) and specificity (82.9%).KeywordsConvolutional neural networksAttentional neural networksBreast cancerHistopathologic imagesConvolutional block attention moduleAttention-augmented convolutionAttention-guided convolutional neural network

BackgroundA colonoscopy can detect colorectal diseases, including cancers, polyps, and inflammatory bowel diseases. A computer-aided diagnosis (CAD) system using deep convolutional neural networks (CNNs) that can recognize anatomical locations during a colonoscopy could efficiently assist practitioners. We aimed to construct a CAD system using a CNN to distinguish colorectal images from parts of the cecum, ascending colon, transverse colon, descending colon, sigmoid colon, and rectum.MethodWe constructed a CNN by training of 9,995 colonoscopy images and tested its performance by 5,121 independent colonoscopy images that were categorized according to seven anatomical locations: the terminal ileum, the cecum, ascending colon to transverse colon, descending colon to sigmoid colon, the rectum, the anus, and indistinguishable parts. We examined images taken during total colonoscopy performed between January 2017 and November 2017 at a single center. We evaluated the concordance between the diagnosis by endoscopists and those by the CNN. The main outcomes of the study were the sensitivity and specificity of the CNN for the anatomical categorization of colonoscopy images.ResultsThe constructed CNN recognized anatomical locations of colonoscopy images with the following areas under the curves: 0.979 for the terminal ileum; 0.940 for the cecum; 0.875 for ascending colon to transverse colon; 0.846 for descending colon to sigmoid colon; 0.835 for the rectum; and 0.992 for the anus. During the test process, the CNN system correctly recognized 66.6% of images.ConclusionWe constructed the new CNN system with clinically relevant performance for recognizing anatomical locations of colonoscopy images, which is the first step in constructing a CAD system that will support us during colonoscopy and provide an assurance of the quality of the colonoscopy procedure.