Optimizing Deep Learning for Computer-Aided Diagnosis of Lung Diseases: An Automated Method Combining Evolutionary Algorithm, Transfer Learning, and Model Compression

Recent developments in Computer Vision have presented novel opportunities to tackle complex healthcare issues,particularly in the field of lung disease diagnosis.One promising avenue involves the use of chest X-Rays,which are commonly utilized in radiology.To fully exploit their potential,researchers have suggested utilizing deep learning methods to construct computer-aided diagnostic systems.However,constructing and compressing these systems presents a significant challenge,as it relies heavily on the expertise of data scientists.To tackle this issue,we propose an automated approach that utilizes an evolutionary algorithm(EA)to optimize the design and compression of a convolutional neural network(CNN)for X-Ray image classification.Our approach accurately classifies radiography images and detects potential chest abnormalities and infections,including COVID-19.Furthermore,our approach incorporates transfer learning,where a pre-trainedCNNmodel on a vast dataset of chest X-Ray images is fine-tuned for the specific task of detecting COVID-19.This method can help reduce the amount of labeled data required for the task and enhance the overall performance of the model.We have validated our method via a series of experiments against state-of-the-art architectures.

Uniaxial Compressive Strength Prediction for Rock Material in Deep Mine Using Boosting-Based Machine Learning Methods and Optimization Algorithms

Traditional laboratory tests for measuring rock uniaxial compressive strength(UCS)are tedious and timeconsuming.There is a pressing need for more effective methods to determine rock UCS,especially in deep mining environments under high in-situ stress.Thus,this study aims to develop an advanced model for predicting the UCS of rockmaterial in deepmining environments by combining three boosting-basedmachine learning methods with four optimization algorithms.For this purpose,the Lead-Zinc mine in Southwest China is considered as the case study.Rock density,P-wave velocity,and point load strength index are used as input variables,and UCS is regarded as the output.Subsequently,twelve hybrid predictive models are obtained.Root mean square error(RMSE),mean absolute error(MAE),coefficient of determination(R2),and the proportion of the mean absolute percentage error less than 20%(A-20)are selected as the evaluation metrics.Experimental results showed that the hybridmodel consisting of the extreme gradient boostingmethod and the artificial bee colony algorithm(XGBoost-ABC)achieved satisfactory results on the training dataset and exhibited the best generalization performance on the testing dataset.The values of R2,A-20,RMSE,and MAE on the training dataset are 0.98,1.0,3.11 MPa,and 2.23MPa,respectively.The highest values of R2 and A-20(0.93 and 0.96),and the smallest RMSE and MAE values of 4.78 MPa and 3.76MPa,are observed on the testing dataset.The proposed hybrid model can be considered a reliable and effective method for predicting rock UCS in deep mines.

基于Douglas-Peucker算法的船舶AIS航迹数据压缩

为解决普通模式下,将海量AIS航迹数据显示在ECDIS平台上效率低、实时性差等问题,设计一种基于DouglasPeucker算法的AIS航迹数据压缩算法。通过分析AIS航迹数据的特征,总结普通模式下ECDIS平台AIS航迹显示实时性差的原因,提出在保留原始航迹特征和误差允许的范围内剔除冗余和重复信息的思想,结合Douglas-Peucker算法,根据设定的不同阈值提取出关键特征点从而对AIS航迹数据进行压缩。在VC2010平台下对该算法进行实现,实践证明,该算法能在较低失真度的前提下对船舶AIS航迹数据进行压缩,提高了轨迹回放、再现效率,与普通模式下ECDIS显示大量AIS航迹相比,系统占用资源少、处理效率高并具有较高的稳定性。

Douglas-Peucker算法在无拓扑矢量数据压缩中的新改进

分析常规Douglas-Peucker算法压缩无拓扑矢量数据时产生公共边"裂缝"现象的原因——公共边被两次或可能更多次压缩,而每次运用Douglas-Peucker算法压缩时所选择的初始点和终点不同造成的。为此,提出公共边对象化Douglas-Peucker改进算法。为实现此算法,首先设计了新的公共边提取算法来提取公共边,然后使用OOP技术,把公共边的相关信息封装成类,最后根据公共边对象提供的信息对多边形的公共边和非公共边分别进行Douglas-Peucker压缩。以广东省行政界线的SVG矢量图为实验对象验证了此算法的有效性,分析了本算法相对于其它Douglas-Peucker改进算法在所需辅助空间和时间效率上的优势。

Douglas-Peucker算法在无拓扑矢量数据压缩中的改进

分析了常规Douglas-Peucker算法压缩无拓扑矢量数据时产生公共边"裂缝"现象的原因,即公共边被两次或可能更多次压缩,而每次运用Douglas-Peucker算法压缩时所选择的初始点和终点不同造成的。为此,提出了公共边对象化Douglas-Peucker改进算法。为实现此算法,首先设计了新的公共边提取算法来提取公共边,然后使用OOP技术,把公共边的相关信息封装成类,最后根据公共边对象提供的信息对多边形的公共边和非公共边分别进行Douglas-Peucker压缩。以广东省行政界线的SVG矢量图为实验对象验证了该算法的有效性,分析了该算法相对于其他Douglas-Peucker改进算法在所需辅助空间和时间效率上的优势。

COMPRESSIBLE VIRTUAL WINDOW ALGORITHM IN PICKING PROCESS CONTROL OF AUTOMATED SORTING SYSTEM

Compared to fixed virtual window algorithm （FVWA）, the dynamic virtual window algorithm （DVWA） determines the length of each virtual container according to the sizes of goods of each order, which saves space of virtual containers and improves the picking efficiency. However, the interval of consecutive goods caused by dispensers on conveyor can not be eliminated by DVWA, which limits a further improvement of picking efficiency. In order to solve this problem, a compressible virtual window algorithm （CVWA） is presented. It not only inherits the merit of DVWA but also compresses the length of virtual containers without congestion of order accumulation by advancing the beginning time of order picking and reasonably coordinating the pace of order accumulation. The simulation result proves that the picking efficiency of automated sorting system is greatly improved by CVWA.

A hybrid quantum encoding algorithm of vector quantization for image compression

Many classical encoding algorithms of vector quantization （VQ） of image compression that can obtain global optimal solution have computational complexity O（N）. A pure quantum VQ encoding algorithm with probability of success near 100% has been proposed, that performs operations 45√N times approximately. In this paper, a hybrid quantum VQ encoding algorithm between the classical method and the quantum algorithm is presented. The number of its operations is less than √N for most images, and it is more efficient than the pure quantum algorithm.

Intelligent modelling of clay compressibility using hybrid meta-heuristic and machine learning algorithms

Compression index Ccis an essential parameter in geotechnical design for which the effectiveness of correlation is still a challenge.This paper suggests a novel modelling approach using machine learning(ML)technique.The performance of five commonly used machine learning(ML)algorithms,i.e.back-propagation neural network(BPNN),extreme learning machine(ELM),support vector machine(SVM),random forest(RF)and evolutionary polynomial regression(EPR)in predicting Cc is comprehensively investigated.A database with a total number of 311 datasets including three input variables,i.e.initial void ratio e0,liquid limit water content wL,plasticity index Ip,and one output variable Cc is first established.Genetic algorithm(GA)is used to optimize the hyper-parameters in five ML algorithms,and the average prediction error for the 10-fold cross-validation(CV)sets is set as thefitness function in the GA for enhancing the robustness of ML models.The results indicate that ML models outperform empirical prediction formulations with lower prediction error.RF yields the lowest error followed by BPNN,ELM,EPR and SVM.If the ranges of input variables in the database are large enough,BPNN and RF models are recommended to predict Cc.Furthermore,if the distribution of input variables is continuous,RF model is the best one.Otherwise,EPR model is recommended if the ranges of input variables are small.The predicted correlations between input and output variables using five ML models show great agreement with the physical explanation.

AN IMPROVED SPARSITY ADAPTIVE MATCHING PURSUIT ALGORITHM FOR COMPRESSIVE SENSING BASED ON REGULARIZED BACKTRACKING

Sparsity Adaptive Matching Pursuit (SAMP) algorithm is a widely used reconstruction algorithm for compressive sensing in the case that the sparsity is unknown. In order to match the sparsity more accurately, we presented an improved SAMP algorithm based on Regularized Backtracking (SAMP-RB). By adapting a regularized backtracking step to SAMP algorithm in each iteration stage, the proposed algorithm can flexibly remove the inappropriate atoms. The experimental results show that SAMP-RB reconstruction algorithm greatly improves SAMP algorithm both in reconstruction quality and computational time. It has better reconstruction efficiency than most of the available matching pursuit algorithms.

Design of quantum VQ iteration and quantum VQ encoding algorithm taking O（√N） steps for data compression

Vector quantization （VQ） is an important data compression method. The key of the encoding of VQ is to find the closest vector among N vectors for a feature vector. Many classical linear search algorithms take O（N） steps of distance computing between two vectors. The quantum VQ iteration and corresponding quantum VQ encoding algorithm that takes O（√N） steps are presented in this paper. The unitary operation of distance computing can be performed on a number of vectors simultaneously because the quantum state exists in a superposition of states. The quantum VQ iteration comprises three oracles, by contrast many quantum algorithms have only one oracle, such as Shor＇s factorization algorithm and Grover＇s algorithm. Entanglement state is generated and used, by contrast the state in Grover＇s algorithm is not an entanglement state. The quantum VQ iteration is a rotation over subspace, by contrast the Grover iteration is a rotation over global space. The quantum VQ iteration extends the Grover iteration to the more complex search that requires more oracles. The method of the quantum VQ iteration is universal.

Design of real-time data compression wireless sensor network based on LZW algorithm

A real-time data compression wireless sensor network based on Lempel-Ziv-Welch encoding(LZW)algorithm is designed for the increasing data volume of terminal nodes when using ZigBee for long-distance wireless communication.The system consists of a terminal node,a router,a coordinator,and an upper computer.The terminal node is responsible for storing and sending the collected data after the LZW compression algorithm is compressed;The router is responsible for the relay of data in the wireless network;The coordinator is responsible for sending the received data to the upper computer.In terms of network function realization,the development and configuration of CC2530 chips on terminal nodes,router nodes,and coordinator nodes are completed using the Z-stack protocol stack,and the network is successfully organized.Through the final simulation analysis and test verification,the system realizes the wireless acquisition and storage of remote data,and reduces the network occupancy rate through the data compression,which has a certain practical value and application prospects.

Efficient implementation of x-ray ghost imaging based on a modified compressive sensing algorithm

Towards efficient implementation of x-ray ghost imaging(XGI),efficient data acquisition and fast image reconstruction together with high image quality are preferred.In view of radiation dose resulted from the incident x-rays,fewer measurements with sufficient signal-to-noise ratio(SNR)are always anticipated.Available methods based on linear and compressive sensing algorithms cannot meet all the requirements simultaneously.In this paper,a method based on a modified compressive sensing algorithm with conjugate gradient descent method(CGDGI)is developed to solve the problems encountered in available XGI methods.Simulation and experiments demonstrate the practicability of CGDGI-based method for the efficient implementation of XGI.The image reconstruction time of sub-second implicates that the proposed method has the potential for real-time XGI.

Lossless embedding: A visually meaningful image encryption algorithm based on hyperchaos and compressive sensing

A novel visually meaningful image encryption algorithm is proposed based on a hyperchaotic system and compressive sensing(CS), which aims to improve the visual security of steganographic image and decrypted quality. First, a dynamic spiral block scrambling is designed to encrypt the sparse matrix generated by performing discrete wavelet transform(DWT)on the plain image. Then, the encrypted image is compressed and quantified to obtain the noise-like cipher image. Then the cipher image is embedded into the alpha channel of the carrier image in portable network graphics(PNG) format to generate the visually meaningful steganographic image. In our scheme, the hyperchaotic Lorenz system controlled by the hash value of plain image is utilized to construct the scrambling matrix, the measurement matrix and the embedding matrix to achieve higher security. In addition, compared with other existing encryption algorithms, the proposed PNG-based embedding method can blindly extract the cipher image, thus effectively reducing the transmission cost and storage space. Finally, the experimental results indicate that the proposed encryption algorithm has very high visual security.

An image joint compression-encryption algorithm based on adaptive arithmetic coding

Through a series of studies on arithmetic coding and arithmetic encryption, a novel image joint compression- encryption algorithm based on adaptive arithmetic coding is proposed. The contexts produced in the process of image compression are modified by keys in order to achieve image joint compression encryption. Combined with the bit-plane coding technique, the discrete wavelet transform coefficients in different resolutions can be encrypted respectively with different keys, so that the resolution selective encryption is realized to meet different application needs. Zero-tree coding is improved, and adaptive arithmetic coding is introduced. Then, the proposed joint compression-encryption algorithm is simulated. The simulation results show that as long as the parameters are selected appropriately, the compression efficiency of proposed image joint compression-encryption algorithm is basically identical to that of the original image compression algorithm, and the security of the proposed algorithm is better than the joint encryption algorithm based on interval splitting.

A Bit-level Text Compression Scheme Based on the ACW Algorithm

This paper presents a description and performance evaluation of a new bit-level, lossless, adaptive, and asymmetric data compression scheme that is based on the adaptive character wordlength （ACW（n）） algorithm. The proposed scheme enhances the compression ratio of the ACW（n） algorithm by dividing the binary sequence into a number of subsequences （s）, each of them satisfying the condition that the number of decimal values （d） of the n-bit length characters is equal to or less than 256. Therefore, the new scheme is referred to as ACW（n, s）, where n is the adaptive character wordlength and s is the number of subsequences. The new scheme was used to compress a number of text files from standard corpora. The obtained results demonstrate that the ACW（n, s） scheme achieves higher compression ratio than many widely used compression algorithms and it achieves a competitive performance compared to state-of-the-art compression tools.

Several Algorithms for Manipulating Compressed Images in JPEG

Currently,in multimedia and image processing technologies, implementing special kinds of image manipulation operations by dealing directly with the compressed image is a work worthy to be concerned with. Theoretical analysis and experiment haVe indicated that some kinds of image processing works can be done very well by dealing with compressed image. In Ans paper, we give some efficient image manipulation operation algorithms operating on the compressed image data. These algorithms have advantages in computing complexity, storage space retirement and image quality.

Degradation algorithm of compressive sensing

The compressive sensing （CS） theory allows people to obtain signal in the frequency much lower than the requested one of sampling theorem. Because the theory is based on the assumption of that the location of sparse values is unknown, it has many constraints in practical applications. In fact, in many cases such as image processing, the location of sparse values is knowable, and CS can degrade to a linear process. In order to take full advantage of the visual information of images, this paper proposes the concept of dimensionality reduction transform matrix and then se- lects sparse values by constructing an accuracy control matrix, so on this basis, a degradation algorithm is designed that the signal can be obtained by the measurements as many as sparse values and reconstructed through a linear process. In comparison with similar methods, the degradation algorithm is effective in reducing the number of sensors and improving operational efficiency. The algorithm is also used to achieve the CS process with the same amount of data as joint photographic exports group （JPEG） compression and acquires the same display effect.

PERFORMANCE COMPARISON OF ECG COMPRESSION ALGORITHMS BASED ON CLINICAL DIAGNOSIS

This paper reviewed the recent progress in the field of electrocardiogram (ECG) compression and compared the efficiency of some compression algorithms. By experimenting on the 500 cases of ECG signals from the ECG database of China, it obtained the numeral indexes for each algorithm. Then by using the automatic diagnostic program developed by Shanghai Zhongshan Hospital, it also got the parameters of the reconstructed signals from linear approximation distance threshold (LADT), wavelet transform (WT), differential pulse code modulation (DPCM) and discrete cosine transform (DCT) algorithm. The results show that when the index of percent of root mean square difference(PRD) is less than 2.5%, the diagnostic agreement ratio is more than 90%; the index of PRD cannot completely show the damage of significant clinical information; the performance of wavelet algorithm exceeds other methods in the same compression ratio (CR). For the statistical result of the parameters of various methods and the clinical diagnostic results, it is of certain value and originality in the field of ECG compression research.

Adaptive block greedy algorithms for receiving multi-narrowband signal in compressive sensing radar reconnaissance receiver

This paper extends the application of compressive sensing(CS) to the radar reconnaissance receiver for receiving the multi-narrowband signal. By combining the concept of the block sparsity, the self-adaption methods, the binary tree search,and the residual monitoring mechanism, two adaptive block greedy algorithms are proposed to achieve a high probability adaptive reconstruction. The use of the block sparsity can greatly improve the efficiency of the support selection and reduce the lower boundary of the sub-sampling rate. Furthermore, the addition of binary tree search and monitoring mechanism with two different supports self-adaption methods overcome the instability caused by the fixed block length while optimizing the recovery of the unknown signal.The simulations and analysis of the adaptive reconstruction ability and theoretical computational complexity are given. Also, we verify the feasibility and effectiveness of the two algorithms by the experiments of receiving multi-narrowband signals on an analogto-information converter(AIC). Finally, an optimum reconstruction characteristic of two algorithms is found to facilitate efficient reception in practical applications.

The Compression Algorithm for the Data Acquisition System in HT-7 Tokamak

HT-7 superconducting tokamak in the Institute of Plasma Physics of the Chinese Academy of Sciences is an experimental device for fusion research in China. The main task of the data acquisition system of HT-7 is to acquire, store, analyze and index the data. The volume of the data is nearly up to hundreds of million bytes. Besides the hardware and software support, a great capacity of data storage, process and transfer is a more important problem. To deal with this problem, the key technology is data compression algorithm. In the paper, the data format in HT-7 is introduced first, then the data compression algorithm, LZO, being a kind of portable lossless data compression algorithm with ANSI C, is analyzed. This compression algorithm, which fits well with the data acquisition and distribution in the nuclear fusion experiment, offers a pretty fast compression and extremely fast decompression. At last the performance evaluation of LZO application in HT-7 is given.