Training of a neural network for image superresolution based on a nonlinear interpolative vector quantizer

Superresolution is the process of extending the spectrum of a diffraction-limited image beyond the optical passband. We consider the neural-network approach to accomplish superresolution and present results on simulated gray-scale images degraded by diffraction blur and additive noise. Images are assumed to be sampled at the Nyquist rate, which requires spatial interpolation for avoiding aliasing, in addition to frequency-domain extrapolation. A novel, to our knowledge, use of vector quantization for the generation of training data sets is also presented. This is accomplished by training of a nonlinear vector quantizer, whose codebooks are subsequently used in the supervised training of the neural network with backpropagation.     

Novel quantum inspired binary neural network algorithm

Sādhanā - In this paper, a quantum based binary neural network algorithm is proposed, named as novel quantum binary neural network algorithm (NQ-BNN). It forms a neural network structure...     

改进型的神经网络PSD非线性补偿研究

为进一步提高PSD传感器线性度,提出了一种基于改进型神经网络(小生境遗传算法和BP算法的混合算法)的PSD传感器的非线性调校方法,该方法利用神经网络良好的非线性映射能力逼近反非线性函数,从而完成非线性校正.仿真结果表明:与BP算法相比,改进型BP神经网络收敛速度快,而且该方法能有效地消除由于制作工艺及技术等因素对传感器非线性的影响.     

BP神经网络的火焰图像温度检测方法

基于炉膛火焰温度不同，与之相对应通过CCD摄取的炉膛火焰图像颜色也不同，提出了一种检测炉膛火焰温度的方法。将获取的炉膛火焰图像RGB模型转换为HSI颜色值，用H、S值作为BP神经网络输入，通过样本图像训练后，拟合H、S与温度T的非线性关系，计算得到炉膛火焰温度值。实验表明，计算温度与实际温度良好相符，温度测量方法切实可行。     

Superresolution of ultrasound images using the first and second harmonic signal

This paper presents a new method of blind two-dimensional (2-D) homomorphic deconvolution and speckle reduction applied to medical ultrasound images. The deconvolution technique is based on an improved 2-D phase unwrapping scheme for pulse estimation. The input images are decomposed into minimum-phase and allpass components. The 2-D phase unwrapping is applied only to the allpass component. The 2-D phase of the minimum-phase component is derived by a Hilbert transform. The accuracy of 2-D phase unwrapping is also improved by processing small (16 x 16 pixels) overlapping subimages separately. This takes the spatial variance of the ultrasound pulse into account. The deconvolution algorithm is applied separately to the first and second harmonic images, producing much sharper images of approximately the same resolution and different speckle patterns. Speckle reduction is made by adding the envelope images of the deconvolved first and second harmonic images. Neither the spatial resolution nor the frame rate decreases, as the common compounding speckle reduction techniques do. The method is tested on sequences of clinical ultrasound images, resulting in high-resolution ultrasound images with reduced speckle noise.     

Multiframe superresolution of binary images

We describe a new algorithm for superresolving a binary object from multiple undersampled low-resolution (LR) images that are degraded by diffraction-limited optical blur, detector blur, and additive white Gaussian noise. Two-dimensional distributed data detection (2D4) is an iterative algorithm that employs a message-passing technique for estimating the object pixel likelihoods. We present a novel non-training-based complexity-reduction technique that makes the algorithm suitable even for channels with support size as large as 5 x 5 object pixels. We compare the performance and computational complexity of 2D4 with that of iterative backprojection (IBP). In an imaging system with an optical blur spot matched to the object pixel size, 2 x 2 undersampled measurement pixels, and four LR images, the reconstruction error measured in terms of the number of pixel mismatches for 2D4 is 300 times smaller than that for IBP at a signal-to-noise ratio of 38 dB.     

Photon-limited image classification with a feedforward neural network

An application of neural networks to the classification of photon-limited images is reported. A three-level feedforward network architecture is employed in which the input units of the network correspond to the pixels of a two-dimensional image. The network is trained in a minicomputer by the use of the backpropagation technique. The statistics of the network components are analyzed, resulting in a method by which the probability of correct classification of a given input image can be calculated. Photon-limited images of printed characters are obtained with a photon-counting camera and are classified. The experimental results are in excellent agreement with theoretical predictions.     

Colored pattern recognition with a neural network model

An algorithm for multicolored pattern recognition is proposed. A scheme for recognizing patterns encoded in three basic colors, i.e., red, green, and blue, is presented. This scheme can be implemented optically with grating structures. Another advantage of this scheme is its capability of pattern recognition with gray levels. This can be accomplished by coding gray levels with different colors.     

A hyperstable neural network for the modelling and control of nonlinear systems

A multivariable hyperstable robust adaptive decoupling control algorithm based on a neural network is presented for the control of nonlinear multivariable coupled systems with unknown parameters and structure. The Popov theorem is used in the design of the controller. The modelling errors, coupling action and other uncertainties of the system are identified on-line by a neural network. The identified results are taken as compensation signals such that the robust adaptive control of nonlinear systems is realised. Simulation results are given.     

基于人工神经网络的传感器非线性拟合方法的研究

人工神经网络具有以任意精度逼近任何非线性连续函数的能力。文章分别用BP网络、RBF网络对传感器的非线性作高精度拟合。仿真结果表明,三层BP网络和RBF网络能够满足工程实际中一维数据拟合的要求,网络具有良好的泛化能力。     

Quantitative measurement of texture orientation in biomedical images using an artificial neural network

Texture orientation is one of the most important attributes used in biomedical and clinical image interpretation. It provides critical clues of continuity and connectivity useful in relating adjacent image areas. We report a novel approach in which image data are convolved with directional convolution masks and the results are used as input to an artificial neural network for classification of image areas into a number of discrete texture orientation classes. © 1998 John Wiley & Sons, Inc. Int J Imaging Syst Technol, 9, 351–355, 1998     

Superresolution with an apodization film in a confocal setup

We put forward the idea of implementing a confocal setup to suppress the large sidelobes spreading all through the field outside the central core. They are produced when an apodization film is imposed on an ordinary lens. Furthermore, this avoids the imaging quality degradation caused by nonaxial points and non-Gaussian-plane points. Also a configuration for achieving three-dimensional superresolution is depicted and discussed.     

Optical-mode neural network by use of the nonlinear response of a laser diode to external optical feedback

We present an intelligent all-optical neural network using a single laser diode that is provided with controlled external feedback. The outputs of the laser neural network (LNN) are represented in the optical domain by the longitudinal cavity modes of the laser diode. The inputs to the LNN are applied by means of adjusting the external feedback of each longitudinal mode through an optical vector-matrix multiplier. Supervised training of some basic input-output mappings is demonstrated by means of a stochastic learning algorithm. The stability and reproducibility of the LNN setup is examined.     

Analysis and classification of malignancy in pancreatic magnetic resonance images using neural network techniques

Computer-aided diagnosis (CAD) is a computerized way of detecting tumors in MR images. Magnetic resonance imaging (MRI) has been generally used in the diagnosis and detection of pancreatic tumors. In a medical imaging system, soft tissue contrast and noninvasiveness are clear preferences of MRI. Inaccurate detection of tumor and long time consumption are the disadvantages of MRI. Computerized classifiers can greatly renew the diagnosis activity, in terms of both accuracy and time necessity by normal and abnormal images, automatically. This article presents an intelligent, automatic, accurate, and robust method to classify human pancreas MRI images as normal or abnormal in terms of pancreatic tumor. It represents the response of artificial neural network (ANN) and support vector machine (SVM) techniques for pancreatic tumor classification. For this, we extract features from MR images of pancreas using the GLCM method and select the best features using JAFER algorithm. These features are analyzed by five classification techniques: ANN BP, ANN RBF, SVM Linear, SVM Poly, and SVM RBF. We compare the results with benchmark data set of MR brain images. The analytical outcome presents that the two best features used to classify the MR images using ANN BP technique have 98% classification accuracy.     

Thyroid nodules classification and diagnosis in ultrasound images using fine-tuning deep convolutional neural network

Ultrasonography AKA diagnostic sonography is a noninvasive imaging technique that allows the analysis of an organic structure, thanks to the ultrasonic waves. It is a valuable diagnosis method and is also seen as the evidence-based diagnostic method for thyroid nodules. The diagnosis, however, is visually made by the practitioner. The automatic discrimination of benign and malignant nodules would be very useful to report Thyroid Imaging Reporting. In this paper, we propose a fine-tuning approach based on deep learning using a Convolutional Neural Network model named resNet-50. This approach allows improving the effectiveness of the classification of thyroid nodules in ultrasound images. Experiments have been conducted on 814 ultrasound images and the results show that our proposed approach dramatically improves the accuracy of the classification of thyroid nodules and outperforms The VGG-19 model.     

Optical properties of binary composite materials with two nonlinear components

Abstract

A new formulation is presented for the calculation of effective dielectric magnitudes of two-component composites in which both components (the host matrix particles and the embedded particles) exhibit nonlinear behaviour of the Kerr type. It is predicted that, under certain conditions, two nonlinear component composites can exhibit optical bistable behaviour as a function of the shape and concentration of the embedded particles, the dielectric contants of the components, the intensity of the external electric field (power density) and the intrinsic third-order nonlinear optical susceptibilities χ⁽³⁾ _p and χ⁽³⁾ _m of the nonlinear components. It is also deduced that, as the power density increases, the effective third-order nonlinear optical susceptibility χ⁽³⁾ of the composite exhibits a clear transition from values close to χ⁽³⁾ _p (low power density) to χ⁽³⁾ _m (high power density). Therefore, it is shown that the optical response of binary composites dramatically changes at moderate and high power densities. A comparison is performed between the optical response of a real two nonlinear component composite and that of a composite with a single nonlinear component.     

Loop mirror laser neural network with a fast liquid-crystal display

In our laser neural network (LNN) all-optical threshold action is obtained by application of controlled optical feedback to a laser diode. Here an extended experimental LNN is presented with as many as 32 neurons and 12 inputs. In the setup we use a fast liquid-crystal display to implement an optical matrix vector multiplier. This display, based on ferroelectric liquid-crystal material, enables us to present 125 training examples/s to the LNN. To maximize the optical feedback efficiency of the setup, a loop mirror is introduced. We use a delta-rule learning algorithm to train the network to perform a number of functions toward the application area of telecommunication data switching.     

Breast cancer diagnosis from mammographic images using optimized feature selection and neural network architecture

Breast cancer is one of the deadly diseases in women that have raised the mortality rate of women. An accurate and early detection of breast cancer using mammogram images is still a complex task. Hence, this article proposes a novel breast cancer detection model, which included five major phases: (a) preprocessing, (b) segmentation, (c) feature extraction, (d) feature selection, and (e) classification. The input mammogram image is initially preprocessed using contrast limited adaptive histogram equalization (CLAHE) and median filtering. The preprocessed image is then subjected to segmentation via the region growing algorithm. Subsequently, geometric features, texture features and gradient features are extracted from the segmented image. Since the length of the feature vector is large, it is essential to select the optimal features. Here, the selection of optimal features is done by a hybrid optimization algorithm. Once the optimal features are selected, they are subjected to the classification process involving the neural network (NN) classifier. As a novelty, the weight of NN is selected optimally to enhance the accuracy of diagnosis (benign and malignant). The optimal feature selection as well as the weight optimization of NN is accomplished by merging the Lion algorithm (LA) and particle swarm optimization (PSO), named as velocity updated lion algorithm (VU‐LA). Finally, a performance‐based evaluation is carried out between VU‐LA and the existing models like, whale optimization algorithm (WOA), gray wolf optimization (GWO), firefly (FF), PSO, and LA.     

Optical learning neural network with a pockels readout optical modulator

We have constructed an optical neural-network system with learning capability by using a Pockels readout optical modulator. The system has a two-dimensional structure that permits easy optical alignment and can handle images without scanning. Learning signals are calculated optically with two liquid-crystal devices by a matrix-matrix outer-product method. The calculated learning signals are added directly to the weights memorized on the Pockels readout optical modulator. A two-layer network is implemented, and the learning and recognition of four alphabetical characters are realized according to the delta rule.     

Using a convolutional neural network model to derive imaging landmarks for lumbar spine numbering on axial magnetic resonance images

Understanding the axial lumbar spine anatomy, including knowledge of the relationship between the lumbar spine level and other paraspinal structures, is important for diagnosing and treating diseases. The purpose of this study was to validate the accuracy of a convolutional neural network (CNN) model in lumbar spine level numbering on axial magnetic resonance (MR) images and to find the appropriate anatomic landmarks for numbering using a class activation map (CAM). A total of 6055 axial MR images of the lumbar spine from the L1-2 to L5-S1 disc levels were obtained to train and validate the CNN model. MR images were acquired using three 3-Tesla machines. The algorithm was developed with three models, and the best-performing model was selected. The external validation set (n = 493) was obtained from other institutions using various machines. The accuracy of the numbering was analyzed using a confusion matrix and receiver operating characteristic curves. The CAMs were reviewed, and the identified anatomic structures were investigated. A reader study was performed by three radiologists, and their accuracy was compared with that of the model. The overall accuracy of the best-performing model for lumbar spine numbering was 0.98 on internal validation and 0.95 on external validation. For the CAM review, mappings concentrated on both paraspinal areas, including the kidney, back muscles, and ilium according to the level. Top-1 and top-2 accuracies of the reviewers ranged between 0.56–0.75, and 0.84–0.93, respectively. After reviewing the CAMs, the accuracy increased to 0.75–0.78 and 0.93–0.98, respectively. A CNN model can accurately determine the level of the lumbar spine on axial MR images, and the configuration of muscles can be used to determine the lumbar level.