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.

A Railway Fastener Inspection Method Based on Abnormal Sample Generation

Regular fastener detection is necessary to ensure the safety of railways.However,the number of abnormal fasteners is significantly lower than the number of normal fasteners in real railways.Existing supervised inspectionmethods have insufficient detection ability in cases of imbalanced samples.To solve this problem,we propose an approach based on deep convolutional neural networks(DCNNs),which consists of three stages:fastener localization,abnormal fastener sample generation based on saliency detection,and fastener state inspection.First,a lightweight YOLOv5s is designed to achieve fast and precise localization of fastener regions.Then,the foreground clip region of a fastener image is extracted by the designed fastener saliency detection network(F-SDNet),combined with data augmentation to generate a large number of abnormal fastener samples and balance the number of abnormal and normal samples.Finally,a fastener inspection model called Fastener ResNet-8 is constructed by being trained with the augmented fastener dataset.Results show the effectiveness of our proposed method in solving the problem of sample imbalance in fastener detection.Qualitative and quantitative comparisons show that the proposed F-SDNet outperforms other state-of-the-art methods in clip region extraction,reaching MAE and max F-measure of 0.0215 and 0.9635,respectively.In addition,the FPS of the fastener state inspection model reached 86.2,and the average accuracy reached 98.7%on 614 augmented fastener test sets and 99.9%on 7505 real fastener datasets.

Calibration of CO and CO2 Monitors Used in Periodic Inspection of Vehicles at Fixed Stations for Environmental Control

Global efforts for environmental cleanliness through the control of gaseous emissions from vehicles are gaining momentum and attracting increasing attention. Calibration plays a crucial role in these efforts by ensuring the quantitative assessment of emissions for informed decisions on environmental treatments. This paper describes a method for the calibration of CO/CO2 monitors used for periodic inspections of vehicles in cites. The calibration was performed in the selected ranges: 900 - 12,000 µmol/mol for CO and 2000 - 20,000 µmol/mol for CO2. The traceability of the measurement results to the SI units was ensured by using certified reference materials from CO/N2 and CO2/N2 primary gas mixtures. The method performance was evaluated by assessing its linearity, accuracy, precision, bias, and uncertainty of the calibration results. The calibration data exhibited a strong linear trend with R² values close to 1, indicating an excellent fit between the measured values and the calibration lines. Precision, expressed as relative standard deviation (%RSD), ranged from 0.48 to 4.56% for CO and from 0.97 to 3.53% for CO2, staying well below the 5% threshold for reporting results at a 95% confidence level. Accuracy measured as percent recovery, was consistently high (≥ 99.1%) for CO and ranged from 84.90% to 101.54% across the calibration range for CO2. In addition, the method exhibited minimal bias for both CO and CO2 calibrations and thus provided a reliable and accurate approach for calibrating CO/CO2 monitors used in vehicle inspections. Thus, it ensures the effectiveness of exhaust emission control for better environment.

Automatic Road Tunnel Crack Inspection Based on Crack Area Sensing and Multiscale Semantic Segmentation

The detection of crack defects on the walls of road tunnels is a crucial step in the process of ensuring travel safetyand performing routine tunnel maintenance. The automatic and accurate detection of cracks on the surface of roadtunnels is the key to improving the maintenance efficiency of road tunnels. Machine vision technology combinedwith a deep neural network model is an effective means to realize the localization and identification of crackdefects on the surface of road tunnels.We propose a complete set of automatic inspection methods for identifyingcracks on the walls of road tunnels as a solution to the problem of difficulty in identifying cracks during manualmaintenance. First, a set of equipment applied to the real-time acquisition of high-definition images of walls inroad tunnels is designed. Images of walls in road tunnels are acquired based on the designed equipment, whereimages containing crack defects are manually identified and selected. Subsequently, the training and validationsets used to construct the crack inspection model are obtained based on the acquired images, whereas the regionscontaining cracks and the pixels of the cracks are finely labeled. After that, a crack area sensing module is designedbased on the proposed you only look once version 7 model combined with coordinate attention mechanism (CAYOLOV7) network to locate the crack regions in the road tunnel surface images. Only subimages containingcracks are acquired and sent to the multiscale semantic segmentation module for extraction of the pixels to whichthe cracks belong based on the DeepLab V3+ network. The precision and recall of the crack region localizationon the surface of a road tunnel based on our proposed method are 82.4% and 93.8%, respectively. Moreover, themean intersection over union (MIoU) and pixel accuracy (PA) values for achieving pixel-level detection accuracyare 76.84% and 78.29%, respectively. The experimental results on the dataset show that our proposed two-stagedetection method outperforms other state-of-the-art models in crack region localization and detection. Based onour proposedmethod, the images captured on the surface of a road tunnel can complete crack detection at a speed often frames/second, and the detection accuracy can reach 0.25 mm, which meets the requirements for maintenanceof an actual project. The designed CA-YOLO V7 network enables precise localization of the area to which a crackbelongs in images acquired under different environmental and lighting conditions in road tunnels. The improvedDeepLab V3+ network based on lightweighting is able to extract crack morphology in a given region more quicklywhile maintaining segmentation accuracy. The established model combines defect localization and segmentationmodels for the first time, realizing pixel-level defect localization and extraction on the surface of road tunnelsin complex environments, and is capable of determining the actual size of cracks based on the physical coordinatesystemafter camera calibration. The trainedmodelhas highaccuracy andcanbe extendedandapplied to embeddedcomputing devices for the assessment and repair of damaged areas in different types of road tunnels.

Unmanned Aerial Vehicle Inspection Routing and Scheduling for Engineering Management

Technological advancements in unmanned aerial vehicles(UAVs)have revolutionized various industries,enabling the widespread adoption of UAV-based solutions.In engineering management,UAV-based inspection has emerged as a highly efficient method for identifying hidden risks in high-risk construction environments,surpassing traditional inspection techniques.Building on this foundation,this paper delves into the optimization of UAV inspection routing and scheduling,addressing the complexity introduced by factors such as no-fly zones,monitoring-interval time windows,and multiple monitoring rounds.To tackle this challenging problem,we propose a mixed-integer linear programming(MILP)model that optimizes inspection task assignments,monitoring sequence schedules,and charging decisions.The comprehensive consideration of these factors differentiates our problem from conventional vehicle routing problem(VRP),leading to a mathematically intractable model for commercial solvers in the case of large-scale instances.To overcome this limitation,we design a tailored variable neighborhood search(VNS)metaheuristic,customizing the algorithm to efficiently solve our model.Extensive numerical experiments are conducted to validate the efficacy of our proposed algorithm,demonstrating its scalability for both large-scale and real-scale instances.Sensitivity experiments and a case study based on an actual engineering project are also conducted,providing valuable insights for engineering managers to enhance inspection work efficiency.

A Systematic Review of Computer Vision Techniques for Quality Control in End-of-Line Visual Inspection of Antenna Parts

The rapid evolution of wireless communication technologies has underscored the critical role of antennas in ensuring seamless connectivity.Antenna defects,ranging from manufacturing imperfections to environmental wear,pose significant challenges to the reliability and performance of communication systems.This review paper navigates the landscape of antenna defect detection,emphasizing the need for a nuanced understanding of various defect types and the associated challenges in visual detection.This review paper serves as a valuable resource for researchers,engineers,and practitioners engaged in the design and maintenance of communication systems.The insights presented here pave the way for enhanced reliability in antenna systems through targeted defect detection measures.In this study,a comprehensive literature analysis on computer vision algorithms that are employed in end-of-line visual inspection of antenna parts is presented.The PRISMA principles will be followed throughout the review,and its goals are to provide a summary of recent research,identify relevant computer vision techniques,and evaluate how effective these techniques are in discovering defects during inspections.It contains articles from scholarly journals as well as papers presented at conferences up until June 2023.This research utilized search phrases that were relevant,and papers were chosen based on whether or not they met certain inclusion and exclusion criteria.In this study,several different computer vision approaches,such as feature extraction and defect classification,are broken down and analyzed.Additionally,their applicability and performance are discussed.The review highlights the significance of utilizing a wide variety of datasets and measurement criteria.The findings of this study add to the existing body of knowledge and point researchers in the direction of promising new areas of investigation,such as real-time inspection systems and multispectral imaging.This review,on its whole,offers a complete study of computer vision approaches for quality control in antenna parts.It does so by providing helpful insights and drawing attention to areas that require additional exploration.

High-speed railway track components inspection framework based on YOLOv8 with high-performance model deployment

Railway inspection poses significant challenges due to the extensive use of various components in vast railway networks,especially in the case of high-speed railways.These networks demand high maintenance but offer only limited inspection windows.In response,this study focuses on developing a high-performance rail inspection system tailored for high-speed railways and railroads with constrained inspection timeframes.This system leverages the latest artificial intelligence advancements,incorporating YOLOv8 for detection.Our research introduces an efficient model inference pipeline based on a producer-consumer model,effectively utilizing parallel processing and concurrent computing to enhance performance.The deployment of this pipeline,implemented using C++,TensorRT,float16 quantization,and oneTBB,represents a significant departure from traditional sequential processing methods.The results are remarkable,showcasing a substantial increase in processing speed:from 38.93 Frames Per Second(FPS)to 281.06 FPS on a desktop system equipped with an Nvidia RTX A6000 GPU and from 19.50 FPS to 200.26 FPS on the Nvidia Jetson AGX Orin edge computing platform.This proposed framework has the potential to meet the real-time inspection requirements of high-speed railways.

A path planning method for robot patrol inspection in chemical industrial parks

Safety patrol inspection in chemical industrial parks is a complex multi-objective task with multiple degrees of freedom.Traditional pointer instruments with advantages like high reliability and strong adaptability to harsh environment,are widely applied in such parks.However,they rely on manual readings which have problems like heavy patrol workload,high labor cost,high false positives/negatives and poor timeliness.To address the above problems,this study proposes a path planning method for robot patrol in chemical industrial parks,where a path optimization model based on improved iterated local search and random variable neighborhood descent(ILS-RVND)algorithm is established by integrating the actual requirements of patrol tasks in chemical industrial parks.Further,the effectiveness of the model and algorithm is verified by taking real park data as an example.The results show that compared with GA and ILS-RVND,the improved algorithm reduces quantification cost by about 24%and saves patrol time by about 36%.Apart from shortening the patrol time of robots,optimizing their patrol path and reducing their maintenance loss,the proposed algorithm also avoids the untimely patrol of robots and enhances the safety factor of equipment.

From manual to automated fabric inspection

Uster,Switzerland,28th March 2024–Uster Technologies offers a flexible solution to upgrade fabric inspection from manual to automated.Integration in existing production lines is quick and easy,and the data flow also brings extra benefits.It means fabric producers can significantly improve their yield with fast,accurate quality monitoring.

Analysis of the Application of Bridge Rapid Inspection and Evaluation Technology

Rapid bridge inspection and evaluation mainly uses information technology to test the quality of bridge infrastructure and structures,integrates the test results with the existing management system,completes the bridge status assessment,establishes information management files to provide bridge disease problem inspection and analysis,and provides support for the application of disposal measures.This paper briefly discusses the necessity of applying rapid inspection and evaluation technology and analyzes the bridge’s rapid inspection and evaluation content,inspection system,and application process.We look forward to the future application prospects of this technology and supporting those in this field.

Defect Inspection Technology for Steel Truss Suspension Bridges

Steel truss suspension bridges are prone to developing defects after prolonged use.These defects may include corrosion of the main cable or the steel truss.To ensure the normal and safe functioning of the suspension bridge,it is necessary to inspect for defects promptly,understand the cause of the defect,and locate it through the use of inspection technology.By promptly addressing defects,the suspension bridge’s safety can be ensured.The author has analyzed the common defects and causes of steel truss suspension bridges and proposed specific inspection technologies.This research is intended to aid in the timely discovery of steel truss suspension bridge defects.

Surface form inspection with contact coordinate measurement:a review

Parts with high-quality freeform surfaces have been widely used in industries,which require strict quality control during the manufacturing process.Among all the industrial inspection methods,contact measurement with coordinate measuring machines or computer numerical control machine tool is a fundamental technique due to its high accuracy,robustness,and universality.In this paper,the existing research in the contact measurement field is systematically reviewed.First,different configurations of the measuring machines are introduced in detail,which may have influence on the corresponding sampling and inspection path generation criteria.Then,the entire inspection pipeline is divided into two stages,namely the pre-inspection and post-inspection stages.The typical methods of each sub-stage are systematically overviewed and classified,including sampling,accessibility analysis,inspection path generation,probe tip radius compensation,surface reconstruction,and uncertainty analysis.Apart from those classical research,the applications of the emerging deep learning technique in some specific tasks of measurement are introduced.Furthermore,some potential and promising trends are provided for future investigation.

Computer-Aided Diagnosis Model Using Machine Learning for Brain Tumor Detection and Classification

The Brain Tumor(BT)is created by an uncontrollable rise of anomalous cells in brain tissue,and it consists of 2 types of cancers they are malignant and benign tumors.The benevolent BT does not affect the neighbouring healthy and normal tissue;however,the malignant could affect the adjacent brain tissues,which results in death.Initial recognition of BT is highly significant to protecting the patient’s life.Generally,the BT can be identified through the magnetic resonance imaging(MRI)scanning technique.But the radiotherapists are not offering effective tumor segmentation in MRI images because of the position and unequal shape of the tumor in the brain.Recently,ML has prevailed against standard image processing techniques.Several studies denote the superiority of machine learning(ML)techniques over standard techniques.Therefore,this study develops novel brain tumor detection and classification model using met heuristic optimization with machine learning(BTDC-MOML)model.To accomplish the detection of brain tumor effectively,a Computer-Aided Design(CAD)model using Machine Learning(ML)technique is proposed in this research manuscript.Initially,the input image pre-processing is performed using Gaborfiltering(GF)based noise removal,contrast enhancement,and skull stripping.Next,mayfly optimization with the Kapur’s thresholding based segmentation process takes place.For feature extraction proposes,local diagonal extreme patterns(LDEP)are exploited.At last,the Extreme Gradient Boosting(XGBoost)model can be used for the BT classification process.The accuracy analysis is performed in terms of Learning accuracy,and the validation accuracy is performed to determine the efficiency of the proposed research work.The experimental validation of the proposed model demonstrates its promising performance over other existing methods.

Computer-Aided Diagnosis for Tuberculosis Classification with Water Strider Optimization Algorithm

Computer-aided diagnosis(CAD)models exploit artificial intelligence(AI)for chest X-ray(CXR)examination to identify the presence of tuberculosis(TB)and can improve the feasibility and performance of CXR for TB screening and triage.At the same time,CXR interpretation is a time-consuming and subjective process.Furthermore,high resemblance among the radiological patterns of TB and other lung diseases can result in misdiagnosis.Therefore,computer-aided diagnosis(CAD)models using machine learning(ML)and deep learning(DL)can be designed for screening TB accurately.With this motivation,this article develops a Water Strider Optimization with Deep Transfer Learning Enabled Tuberculosis Classification(WSODTL-TBC)model on Chest X-rays(CXR).The presented WSODTL-TBC model aims to detect and classify TB on CXR images.Primarily,the WSODTL-TBC model undergoes image filtering techniques to discard the noise content and U-Net-based image segmentation.Besides,a pre-trained residual network with a two-dimensional convolutional neural network(2D-CNN)model is applied to extract feature vectors.In addition,the WSO algorithm with long short-term memory(LSTM)model was employed for identifying and classifying TB,where the WSO algorithm is applied as a hyperparameter optimizer of the LSTM methodology,showing the novelty of the work.The performance validation of the presented WSODTL-TBC model is carried out on the benchmark dataset,and the outcomes were investigated in many aspects.The experimental development pointed out the betterment of the WSODTL-TBC model over existing algorithms.

Low-Cost Real-Time Automated Optical Inspection Using Deep Learning and Attention Map

The recent trends in Industry 4.0 and Internet of Things have encour-aged many factory managers to improve inspection processes to achieve automa-tion and high detection rates.However,the corresponding cost results of sample tests are still used for quality control.A low-cost automated optical inspection system that can be integrated with production lines to fully inspect products with-out adjustments is introduced herein.The corresponding mechanism design enables each product to maintain afixed position and orientation during inspec-tion to accelerate the inspection process.The proposed system combines image recognition and deep learning to measure the dimensions of the thread and iden-tify its defects within 20 s,which is lower than the production-line productivity per 30 s.In addition,the system is designed to be used for monitoring production lines and equipment status.The dimensional tolerance of the proposed system reaches 0.012 mm,and a 100%accuracy is achieved in terms of the defect reso-lution.In addition,an attention-based visualization approach is utilized to verify the rationale for the use of the convolutional neural network model and identify the location of thread defects.

Automated X-ray Defect Inspection on Occluded BGA Balls Using Hybrid Algorithm

Automated X-ray defect inspection of occluded objects has been an essential topic in semiconductors,autonomous vehicles,and artificial intelligence devices.However,there are few solutions to segment occluded objects in the X-ray inspection efficiently.In particular,in the Ball Grid Array inspection of X-ray images,it is difficult to accurately segment the regions of occluded solder balls and detect defects inside solder balls.In this paper,we present a novel automatic inspection algorithm that segments solder balls,and detects defects fast and efficiently when solder balls are occluded.The proposed algorithm consists of two stages.In the first stage,the defective candidates or defects are determined through the following four steps:(i)image preprocessing such as noise removal,contrast enhancement,binarization,connected component,and morphology,(ii)limiting the inspec-tion area to the ball regions and determining if the ball regions are occluded,(iii)segmenting each ball region into one or more regions with similar gray values,and(iv)determining whether there are defects or defective candidates in the regions using a weighted sum of local threshold on local variance.If there are defective candidates,the determination of defects is finally made in the following stage.In the second stage,defects are detected using the automated inspection technique based on oblique computed tomography.The 3D precision inspection process is divided into four steps:(i)obtaining 360 projection images(one image per degree)rotating the object from 0 to 360 degrees,(ii)reconstructing a 3D image from the 360 projected images,(iii)finding the center slice of gravity for solder balls from the axial slice images in the z-direction,and getting the inspection intervals between the upper bound and the lower bound from the center slice,and(iv)finally determining whether there are defects in the averaged image of solder balls.The proposed hybrid algorithm is robust for segmenting the defects inside occluded solder balls,and improves the performance of solder ball segmentation and defect detection algorithm.Experimental results show an accuracy of more than 97%.

Defect inspection of indoor components in buildings using deep learning object detection and augmented reality

Visual inspection is commonly adopted for building operation,maintenance,and safety.The durability and defects of components or materials in buildings can be quickly assessed through visual inspection.However,implementations of visual inspection are substantially time-consuming,labor-intensive,and error-prone because useful auxiliary tools that can instantly highlight defects or damage locations from images are not available.Therefore,an advanced building inspection framework is developed and implemented with augmented reality(AR)and real-time damage detection in this study.In this framework,engineers should walk around and film every corner of the building interior to generate the three-dimensional(3D)environment through ARKit.Meanwhile,a trained YOLOv5 model real-time detects defects during this process,even in a large-scale field,and the defect locations indicating the detected defects are then marked in this 3D environment.The defects areas can be measured with centimeter-level accuracy with the light detection and ranging(LiDAR)on devices.All required damage information,including defect positions and sizes,is collected at a time and can be rendered in the 2D and 3D views.Finally,this visual inspection can be efficiently conducted,and the previously generated environment can also be loaded to re-localize existing defect marks for future maintenance and change observation.Moreover,the proposed framework is also implemented and verified by an underground parking lot in a building to detect and quantify surface defects on concrete components.As seen in the results,the conventional building inspection is significantly improved with the aid of the proposed framework in terms of damage localization,damage quantification,and inspection efficiency.

Comparison of three fundus inspection methods during phacoemulsification in diabetic white cataract

AIM:To investigate whether Wild Field Imaging System(WFIS SW-8000),25G endoilluminator,and intraoperative optical coherence tomography(iOCT)can perform realtime screening and diagnosing in patients with suspicious diabetic retinopathy(DR)during phacoemulsification,especially in cases of white cataract.METHODS:A cross-sectional study was carried out.A total of 204 dense diabetic cataractous eyes of 204 patients with suspected DR treated from April 2020 to March 2021 were included.Phacoemulsification combined with intraocular lens implantation was performed.Following the removal of the lens opacity,the 25G endoilluminator,fundus photography,and iOCT were performed successively.Optical coherence tomography(OCT)and/or fundus fluorescein angiography(FFA)were used to verify the fundus findings postoperatively.Intraoperative and postoperative results were compared to verify the accuracy of intraoperative diagnosis in each group.RESULTS:Intraoperative and postoperative examinations revealed 58 and 62 eyes with DR,respectively(positive rate,28.43%and 30.39%,respectively).During the phacoemulsification,WFIS SW-8000 detected 44 eyes with DR(the detection rate,70.97%);25G endo-illuminator found 56 eyes with DR(the detection rate,90.32%);iOCT found 46 eyes with DR(the detection rate,74.19%);and 58 eyes with DR were found by combining the three methods(the detection rate,93.55%).There were statistically significant differences in the diagnostic sensitivity for DR among the methods(χ^(2)=16.36,P=0.001).CONCLUSION:WFIS SW-8000,25G endo-illuminator,iOCT,and especially their combination can be used to inspect the fundus and detect DR intraoperatively;they are helpful for the timely diagnosis and treatment of DR in patients with dense cataract.

Novel Computer-Aided Diagnosis System for the Early Detection of Alzheimer’s Disease

Aging is a natural process that leads to debility,disease,and dependency.Alzheimer’s disease(AD)causes degeneration of the brain cells leading to cognitive decline and memory loss,as well as dependence on others to fulfill basic daily needs.AD is the major cause of dementia.Computer-aided diagnosis(CADx)tools aid medical practitioners in accurately identifying diseases such as AD in patients.This study aimed to develop a CADx tool for the early detection of AD using the Intelligent Water Drop(IWD)algorithm and the Random Forest(RF)classifier.The IWD algorithm an efficient feature selection method,was used to identify the most deterministic features of AD in the dataset.RF is an ensemble method that leverages multiple weak learners to classify a patient’s disease as either demented(DN)or cognitively normal(CN).The proposed tool also classifies patients as mild cognitive impairment(MCI)or CN.The dataset on which the performance of the proposed CADx was evaluated was sourced from the Alzheimer’s Disease Neuroimaging Initiative(ADNI).The RF ensemble method achieves 100%accuracy in identifying DN patients from CN patients.The classification accuracy for classifying patients as MCI or CN is 92%.This study emphasizes the significance of pre-processing prior to classification to improve the classification results of the proposed CADx tool.

Probability Distributions Arising in Connection with the Inspection Paradox for Bernoulli Trials

In renewal theory, the Inspection Paradox refers to the fact that an interarrival period in a renewal process which contains a fixed inspection time tends to be longer than one for the corresponding uninspected process. We focus on the paradox for Bernoulli trials. Probability distributions and moments for the lengths of the interarrival periods are derived for the inspected process, and we compare them to those for the uninspected case.