A Performance Evaluation of Classic Convolutional Neural Networks for 2D and 3D Palmprint and Palm Vein Recognition

Palmprint recognition and palm vein recognition are two emerging biometrics technologies. In the past two decades, many traditional methods have been proposed for palmprint recognition and palm vein recognition, and have achieved impressive results. However, the research on deep learning-based palmprint recognition and palm vein recognition is still very preliminary. In this paper, in order to investigate the problem of deep learning based 2D and 3D palmprint recognition and palm vein recognition in-depth, we conduct performance evaluation of seventeen representative and classic convolutional neural networks (CNNs) on one 3D palmprint database, five 2D palmprint databases and two palm vein databases. A lot of experiments have been carried out in the conditions of different network structures, different learning rates, and different numbers of network layers. We have also conducted experiments on both separate data mode and mixed data mode. Experimental results show that these classic CNNs can achieve promising recognition results, and the recognition performance of recently proposed CNNs is better. Particularly, among classic CNNs, one of the recently proposed classic CNNs, i.e., EfficientNet achieves the best recognition accuracy. However, the recognition performance of classic CNNs is still slightly worse than that of some traditional recognition methods.

     

COVID-19 classification by CCSHNet with deep fusion using transfer learning and discriminant correlation analysis

Aim: COVID-19 is a disease caused by a new strain of coronavirus. Up to 18th October 2020, worldwide there have been 39.6 million confirmed cases resulting in more than 1.1 million deaths. To improve diagnosis, we aimed to design and develop a novel advanced AI system for COVID-19 classification based on chest CT (CCT) images.Methods: Our dataset from local hospitals consisted of 284 COVID-19 images, 281 community-acquired pneumonia images, 293 secondary pulmonary tuberculosis images; and 306 healthy control images. We first used pretrained models (PTMs) to learn features, and proposed a novel (L, 2) transfer feature learning algorithm to extract features, with a hyperparameter of number of layers to be removed (NLR, symbolized as L). Second, we proposed a selection algorithm of pretrained network for fusion to determine the best two models characterized by PTM and NLR. Third, deep CCT fusion by discriminant correlation analysis was proposed to help fuse the two features from the two models. Micro-averaged (MA) F1 score was used as the measuring indicator. The final determined model was named CCSHNet.Results: On the test set, CCSHNet achieved sensitivities of four classes of 95.61%, 96.25%, 98.30%, and 97.86%, respectively. The precision values of four classes were 97.32%, 96.42%, 96.99%, and 97.38%, respectively. The F1 scores of four classes were 96.46%, 96.33%, 97.64%, and 97.62%, respectively. The MA F1 score was 97.04%. In addition, CCSHNet outperformed 12 state-of-the-art COVID-19 detection methods.Conclusions: CCSHNet is effective in detecting COVID-19 and other lung infectious diseases using first-line clinical imaging and can therefore assist radiologists in making accurate diagnoses based on CCTs.     

Performance evaluation of anti-radiation based on the gamma degradation process

In the environment of space radiation, the high-energy charged particles or high-energy photons acting on a spacecraft can cause either temporary device degradation or permanent failure. The traditional probability model is difficult to obtain reliable estimation of unit radiation resistance performance with small samples. Considering that different products will change differently after high-energy particle radiation, we construct a model based on the gamma degradation process. This model can efficiently describe the law of unit radiation resistance variation with the total radiation dose levels under the effect of the total dose and displacement damage. Finally, the proposed model is used to assess the anti-radiation performance of the N-channel power MOSFET device STRH60N20FSY3 produced by STM to obtain average unit radiation resistance, survival probability, survival function, etc.     

Mechanical analysis of double-layered circular graphene sheets as building material embedded in an elastic medium

Journal of Central South University - Possessing the unique and highly valuable properties, graphene sheets (GSs) have attracted increasing attention including that from the building engineer due...     

Model checking response times in Networked Automation Systems using jitter bounds

Response time (RT) of Networked Automation Systems (NAS) is affected by timing imperfections induced due to the network, computing and hardware components. Guaranteeing RT in the presence of such timing imperfections is essential for building dependable NAS, and to avoid costly upgrades after deployment in industries.This investigation proposes a methodology and work-flow that combines modelling, simulation, verification, experiments, and software tools to verify the RT of the NAS during the design, rather than after deployment. The RT evaluation work-flow has three phases: model building, modelling and verification. During the model building phase component reaction times are specified and their timing performance is measured by combining experiments with simulation. During the modelling phase, component based mathematical models that capture the network architecture and inter-connection are proposed. Composition of the component models gives the NAS model required for studying the RT performance on system level. Finally, in the verification step, the NAS formal models are abstracted as UPPAAL timed automata with their timing interfaces. To model timing interfaces, the action patterns, and their timing wrapper are proposed. The formal model of high level of abstraction is used to verify the total response time of the NAS where the reactions to be verified are specified using a subset of timed computation tree logic (TCTL) in UPPAAL model checker. The proposed approach is illustrated on an industrial steam boiler deployment.     

Strong Convergence Analysis of Batch Gradient-Based Learning Algorithm for Training Pi-Sigma Network Based on TSK Fuzzy Models

By combining of the benefits of high-order network and TSK (Tagaki-Sugeno-Kang) inference system, Pi-Sigma network is capable to dispose with the nonlinear problems much more effectively, which means it has a compacter construction, and quicker computational speed. The aim of this paper is to present a gradient-based learning method for Pi-Sigma network to train TSK fuzzy inference system. Moreover, some strong convergence results are established based on the weak convergence outcomes, which indicates that the sequence of weighted fuzzy parameters gets to a fixed point. Simulation results show the modified learning algorithm is effective to support the theoretical results.     

Image inpainting algorithm based on TV model and evolutionary algorithm

With the development of modern image processing techniques, the numbers of images increase at a high speed in network. As a new form of visual communication, image is widely used in network transmission. However, the image information would be lost after transmission. In view of this, we are motivated to restore the image to make it complete in an effective and efficient way in order to save the network bandwidth. At present, there are two main methods for digital image restoration, texture-based method and non-textured-based method. In the texture-based method, Criminisi algorithm is a widely used algorithm. However, the inaccurate completion order and the inefficiency in searching matching patches are two main limitations of Criminisi algorithm. To overcome these shortcomings, in this paper, an exemplar image completion based on evolutionary algorithm is proposed. In the non-textured-based method, total variation method is a typical algorithm. An improved total variation algorithm is proposed in this paper. In the improved algorithm, the diffusion coefficients are defined according to the distance and direction between the damaged pixel and its neighborhood pixel. Experimental results show that the proposed algorithms have better general performance in image completion. And these two new algorithms could improve the experience of network surfing and reduce the network communication cost.     

Positive <Emphasis Type="Italic">L</Emphasis><Subscript>1</Subscript>-gain filter design for positive continuous-time Markovian jump systems with partly known transition rates

The paper is concerned with the problem of positive L ₁-gain filter design for positive continuous-time Markovian jump systems with partly known transition rates. Our aim is to design a positive full-order filter such that the corresponding filtering error system is positive and stochastically stable with L ₁-gain performance. By applying a linear co-positive Lyapunov function and free-connection weighting vectors, the desired positive L ₁-gain filter is provided. The obtained theoretical results are demonstrated by numerical examples.