Connected Components-based Colour Image Representations of Vibrations for a Two-stage Fault Diagnosis of Roller Bearings Using Convolutional Neural Networks

Roller bearing failure is one of the most common faults in rotating machines.Various techniques for bearing fault diagnosis based on faults feature extraction have been proposed.But feature extraction from fault signals requires expert prior information and human labour.Recently,deep learning algorithms have been applied extensively in the condition monitoring of rotating machines to learn features automatically from the input data.Given its robust performance in image recognition,the convolutional neural network(CNN)architecture has been widely used to learn automatically discriminative features from vibration images and classify health conditions.This paper proposes and evaluates a two-stage method RGBVI-CNN for roller bearings fault diagnosis.The first stage in the proposed method is to generate the RGB vibration images(RGBVIs)from the input vibration signals.To begin this process,first,the 1-D vibration signals were converted to 2-D grayscale vibration Images.Once the conversion was completed,the regions of interest(ROI)were found in the converted 2-D grayscale vibration images.Finally,to produce vibration images with more discriminative characteristics,an algorithm was applied to the 2-D grayscale vibration images to produce connected components-based RGB vibration images(RGBVIs)with sets of colours and texture features.In the second stage,with these RGBVIs a CNN-based architecture was employed to learn automatically features from the RGBVIs and to classify bearing health conditions.Two cases of fault classification of rolling element bearings are used to validate the proposed method.Experimental results of this investigation demonstrate that RGBVI-CNN can generate advantageous health condition features from bearing vibration signals and classify the health conditions under different working loads with high accuracy.Moreover,several classification models trained using RGBVI-CNN offered high performance in the testing results of the overall classification accuracy,precision,recall,and F-score.

Design of a novel hybrid quantum deep neural network in INEQR images classification

We redesign the parameterized quantum circuit in the quantum deep neural network, construct a three-layer structure as the hidden layer, and then use classical optimization algorithms to train the parameterized quantum circuit, thereby propose a novel hybrid quantum deep neural network(HQDNN) used for image classification. After bilinear interpolation reduces the original image to a suitable size, an improved novel enhanced quantum representation(INEQR) is used to encode it into quantum states as the input of the HQDNN. Multi-layer parameterized quantum circuits are used as the main structure to implement feature extraction and classification. The output results of parameterized quantum circuits are converted into classical data through quantum measurements and then optimized on a classical computer. To verify the performance of the HQDNN, we conduct binary classification and three classification experiments on the MNIST(Modified National Institute of Standards and Technology) data set. In the first binary classification, the accuracy of 0 and 4 exceeds98%. Then we compare the performance of three classification with other algorithms, the results on two datasets show that the classification accuracy is higher than that of quantum deep neural network and general quantum convolutional neural network.

Sparse representation scheme with enhanced medium pixel intensity for face recognition

Sparse representation is an effective data classification algorithm that depends on the known training samples to categorise the test sample.It has been widely used in various image classification tasks.Sparseness in sparse representation means that only a few of instances selected from all training samples can effectively convey the essential class-specific information of the test sample,which is very important for classification.For deformable images such as human faces,pixels at the same location of different images of the same subject usually have different intensities.Therefore,extracting features and correctly classifying such deformable objects is very hard.Moreover,the lighting,attitude and occlusion cause more difficulty.Considering the problems and challenges listed above,a novel image representation and classification algorithm is proposed.First,the authors’algorithm generates virtual samples by a non-linear variation method.This method can effectively extract the low-frequency information of space-domain features of the original image,which is very useful for representing deformable objects.The combination of the original and virtual samples is more beneficial to improve the clas-sification performance and robustness of the algorithm.Thereby,the authors’algorithm calculates the expression coefficients of the original and virtual samples separately using the sparse representation principle and obtains the final score by a designed efficient score fusion scheme.The weighting coefficients in the score fusion scheme are set entirely automatically.Finally,the algorithm classifies the samples based on the final scores.The experimental results show that our method performs better classification than conventional sparse representation algorithms.

Representation of the Image of Modern China in the Perspective of the Purpose of Peregrination in China

Since the middle and late 18th century,the purpose of western peregrinators to China had been in a gradual process from geological survey,to commercial route exploration,commercial development,and then the struggle to divide power and sphere of influence.The purpose,mode,theme and narrative strategy of missionaries,scientific explorers,business travelers and pilgrims traveling in China were the main elements that affected the representation,construction and production of the image of modern China,as well as the main factors that formed the differentiation of the image of China.

A novel binary image representation algorithm by using NAM and coordinate encoding procedure and its application to area calculation

We propose a novel binary image representation algorithm using the non-symmetry and anti-packing model and the coordinate encoding procedure （NAMCEP）. By tak- ing some idiomatic standard binary images in the field of image processing as typical test objects, and by comparing our proposed NAMCEP representation with linear quadtree （LQT）, binary tree （Bintree）, non-symmetry and anti-packing model （NAM） with K-lines （NAMK）, and NAM representa- tions, we show that NAMCEP can not only reduce the aver- age node, but also simultaneously improve the average com- pression. We also present a novel NAMCEP-based algorithm for area calculation and show experimentally that our algo- rithm offers significant improvements.

Class-Driven Non-Negative Matrix Factorization for Image Representation

Non-negative matrix fa^torization （NMF） is a useful technique to learn a parts-based representation by decom- posing the original data matrix into a basis set and coefficients with non-negative constraints. However, as an unsupervised method, the original NMF cannot utilize the discriminative class information. In this paper, we propose a semi-supervised class-driven NMF method to associate a class label with each basis vector by introducing an inhomogeneous representation cost constraint. This constraint forces the learned basis vectors to represent better for their own classes but worse for the others. Therefore, data samples in the same class will have similar representations, and consequently the discriminability in new representations could be boosted. Some experiments carried out on several standard databases validate the effectiveness of our method in comparison with the state-of-the-art approaches.

Quantum color image scaling based on bilinear interpolation

As a part of quantum image processing, quantum image scaling is a significant technology for the development of quantum computation. At present, most of the quantum image scaling schemes are based on grayscale images, with relatively little processing for color images. This paper proposes a quantum color image scaling scheme based on bilinear interpolation, which realizes the 2^(n_(1)) × 2^(n_(2)) quantum color image scaling. Firstly, the improved novel quantum representation of color digital images(INCQI) is employed to represent a 2^(n_(1)) × 2^(n_(2)) quantum color image, and the bilinear interpolation method for calculating pixel values of the interpolated image is presented. Then the quantum color image scaling-up and scaling-down circuits are designed by utilizing a series of quantum modules, and the complexity of the circuits is analyzed.Finally, the experimental simulation results of MATLAB based on the classical computer are given. The ultimate results demonstrate that the complexities of the scaling-up and scaling-down schemes are quadratic and linear, respectively, which are much lower than the cubic function and exponential function of other bilinear interpolation schemes.

Learning Noise-Assisted Robust Image Features for Fine-Grained Image Retrieval

Fine-grained image search is one of the most challenging tasks in computer vision that aims to retrieve similar images at the fine-grained level for a given query image.The key objective is to learn discriminative fine-grained features by training deep models such that similar images are clustered,and dissimilar images are separated in the low embedding space.Previous works primarily focused on defining local structure loss functions like triplet loss,pairwise loss,etc.However,training via these approaches takes a long training time,and they have poor accuracy.Additionally,representations learned through it tend to tighten up in the embedded space and lose generalizability to unseen classes.This paper proposes a noise-assisted representation learning method for fine-grained image retrieval to mitigate these issues.In the proposed work,class manifold learning is performed in which positive pairs are created with noise insertion operation instead of tightening class clusters.And other instances are treated as negatives within the same cluster.Then a loss function is defined to penalize when the distance between instances of the same class becomes too small relative to the noise pair in that class in embedded space.The proposed approach is validated on CARS-196 and CUB-200 datasets and achieved better retrieval results(85.38%recall@1 for CARS-196%and 70.13%recall@1 for CUB-200)compared to other existing methods.

Integrating Color and Spatial Feature for Content-Based Image Retrieval

In this paper, we present a novel and efficient scheme for extracting, indexing and retrieving color images. Our motivation was to reduce the space overhead of partition-based approaches taking advantage of the fact that only a relatively low number of distinct values of a particular visual feature is present in most images. To extract color feature and build indices into our image database we take into consideration factors such as human color perception and perceptual range, and the image is partitioned into a set of regions by using a simple classifying scheme. The compact color feature vector and the spatial color histogram, which are extracted from the seqmented image region, are used for representing the color and spatial information in the image. We have also developed the region-based distance measures to compare the similarity of two images. Extensive tests on a large image collection were conducted to demonstrate the effectiveness of the proposed approach.

Fast Object Extraction and Euler Number on Block Represented Images

The identification of objects in binary images is a fundamental task in image analysis and pattern recognition tasks. The Euler number of a binary image is an important topological measure which is used as a feature in image analysis. In this paper, a very fast algorithm for the detection and localization of the objects and the computation of the Euler number of a binary image is proposed. The proposed algorithm operates in one scan of the image and is based on the Image Block Representation (IBR) scheme. The proposed algorithm is more efficient than conventional pixel based algorithms in terms of execution speed and representation of the extracted information.

A NAM Representation Method for Data Compression of Binary Images

A representation method using the non-symmetry and anti-packing model （NAM） for data compression of binary images is presented. The NAM representation algorithm is compared with the popular linear quadtree and run length encoding algorithms. Theoretical and experimental results show that the algorithm has a higher compression ratio for both Iossy and Iossless cases of binary images and better reconstructed quality for the Iossy case.

基于非对称逆布局表示的图像周长计算方法(英文)

The non-symmetry anti-packing image representation （NAIR） uses a sequence of the instances of some predefined prototypes to represent an image. While significantly reducing the instances required to represent an image in contrary to the quadtree and the linear quadtree, however, NAIR has lost the explicit space relationship among these instances and hence made some geometric operations such as perimeter computation hard to be implemented. In this paper, longitude and latitude grid （L^2G）, a data structure which can restore lost space relationship from the NAIR is first presented, and then a novel algorithm to compute the perimeters of the images represented by the NAIR is presented. The experimental results show that the new algorithm has saved at least 90% of the running time comparing with that based on the quadtree.

A fast algorithm for computing moments of gray images based on NAM and extended shading approach

Computing moments on images is very important in the fields of image processing and pattern recognition. The non-symmetry and anti-packing model （NAM） is a general pattern representation model that has been developed to help design some efficient image representation methods. In this paper, inspired by the idea of computing moments based on the S-Tree coding （STC） representation and by using the NAM and extended shading （NAMES） approach, we propose a fast algorithm for computing lower order moments based on the NAMES representation, which takes O（N） time where N is the number of NAM blocks. By taking three idiomatic standard gray images ＇Lena＇, ＇F16＇, and ＇Peppers＇ in tile field of image processing as typical test objects, and by comparing our proposed algorithm with the conventional algorithm and the popular STC representation algorithm for computing the lower order moments, the theoretical and experimental results presented in this paper show that the average execution time improvement ratios of the proposed NAMES approach over the STC approach, and also the conventional approach are 26.63%, and 82.57% respectively while maintaining the image quality.

Continuous representation of human portraits and natural scenery in human ventral temporal cortex: evidence from functional magnetic resonance imaging

Background Functional magnetic resonance imaging (fMRI) has become a powerful tool for tracking human brain activity in vivo. This technique is mainly based on blood oxygenation level dependence (BOLD) contrast. In the present study, we employed this newly developed technique to characterize the neural representations of human portraits and natural sceneries in the human brain.Methods Nine subjects were scanned with a 1.5 T magnetic resonance imaging (MRI) scanner using gradient-recalled echo and echo-planar imaging (GRE-EPI) pulse sequence while they were visually presented with 3 types of white-black photographs: natural scenery, human portraits, and scrambled nonsense pictures. Multiple linear regression was used to identify brain regions responding preferentially to each type of stimulus and common regions for both human portraits and natural scenery. The relative contributions of each type of stimulus to activation in these regions were examined using linear combinations of a general linear test.Results Multiple linear regression analysis revealed two distinct but adjacent regions in both sides of the ventral temporal cortex. The medial region preferentially responded to natural scenery, whereas the lateral one preferentially responded to the human portraits. The general linear test further revealed a distribution gradient such that a change from portraits to scenes shifted areas of activation from lateral to medial.Conclusions The boundary between portrait-associated and scenery-associated areas is not as clear as previously demonstrated. The representations of portraits and scenes in ventral temporal cortex appear to be continuous and overlap.