International Journal of Engineering and Manufacturing @ijem
Journal articles - International Journal of Engineering and Manufacturing
All articles: 649
Braille Recognition using a Camera-enabled Smartphone
Scientific article
The paper proposes a method to process the image of a Braille document that interprets the raised dots on the document and converts them to their equivalent English characters. It was found that under ideal conditions of light and alignment of the Braille document with respect to the smartphone, the application can achieve more than 80% accuracy. The application can be used in the education domain, wherein users who do not understand Braille may help visually-impaired or blind students in their learning activities such as assignments and tests.
Free
Breast Cancer Diagnosis Improvement Based Deep Learning
Scientific article
Background: Globally, Breast cancer is the utmost predominant cancer and it affects millions of women every year. Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI) has been familiar as an efficient modality for diagnosing breast cancer. In spite of DCE-MRI modality being majorly utilized for the classification of breast cancer, the diagnostic performance is still deficient and misclassification occurs. Method: This research proposes the Deep Learning (DL) approach of Dual Attention Deep Convolutional Neural Network (DADCNN) for the classification of breast cancer into two types named benign and malignant. Initially, the DCE-MRI, Rider Breast MRI and Breast MRI datasets are utilized for estimating the effectiveness of the classifier. After collecting the dataset, pre-processing is performed by utilizing data augmentation technique. Then, the augmented data is input for the feature extraction process, which is performed by using DenseNet-121 and ResNet-101 architectures. Then, the extracted features are concatenated by using the feature fusion model and finally, classification is performed to categorize the breast cancer. The DADCNN approach deals with the long input features to selectively focus on the most relevant parts in breast cancer, so it enhances the results. The presented DADCNN approach significantly outperforms the existing methods like MUM-Net-joint prediction, UDFS + SVM, XGBoost, Multivariate Rocket and BI-RADS. The greater accuracy of the proposed DACNN approach suggests DL approach to effectively enhance the classification accuracy in breast cancer. Results: The experimental results establish that the proposed method attains greater results in all performance metrics as compared to the exiting methods like Multi-modality Network (MUM-Net) and Multivariate Rocket algorithm, The suggested DADCNN approach attains the maximum accuracy of 0.931, specificity of 0.924, sensitivity of 0.925, AUC of 0.962, PPV of 0.853 and NPV of 0.902 in breast cancer classification, which denotes that the DACNN effectively classify the cancer into benign and malignant. Concluding Remarks: The DADCNN approach deals with the long input features to selectively focus on the most relevant parts in breast cancer, so it enhances the accuracy, specificity, sensitivity, AUC, PPV and NPV in breast cancer classification.
Free
C2C E-commerce Risk Assessment Based on AHP and Fuzzy Comprehensive Evaluation
Scientific article
Firstly this paper gives an introduction to the current C2C e-commerce transactions in China and gives a review of relevant theories and research. Then, we proposed a new method that AHP and fuzzy comprehensive evaluation method of combining, and discuss the methods used in this risk assessment of the feasibility of e-commerce transactions. Finally, an instance of the application of this method to assess the final results prove that the application of the method of assessment of such problems feasible. With the rapid development of Internet, e-commerce as a new business model has achieved great development and progress which greatly changed the way people live and work, at the same time it gives a rise to the unlimited business opportunities. C2C is one kind of the e-commerce transactions, Internet service providers that use of computer and network technology, used to provide paid or unpaid. E-commerce platform allows two parties (mainly for individual users) in its platform to bid independently, bargaining-based online trading patterns. In recent years, China's C2C market in a rapid growth phase, this model has become the holder of a large number of small venture capital approach. However, due to the virtual nature of online transactions, buyers and sellers transaction process is often non-face environment in, on product quality, price and service and other aspects of information entirely on the buyer's estimate. Therefore, C2C e-commerce system model is a clear system of information asymmetry, the potential risk is enormous. Asymmetric information in this trading environment, to ensure smooth transactions and reduce risk, the two sides of the credit problem is particularly prominent. Meanwhile, in the current transaction before the transaction, the risk prediction and assessment has become very important. Currently, e-commerce e-commerce credit risk has become a constraint to further development of the main obstacles.
Free
CFD Investigation of Thermal and Pressurization Performation in LHe Tank
Scientific article
Research on thermal response of liquid helium storage tank is an important part of non-destructive storage technology for liquid helium, However there are few reports on the thermal response of liquid helium storage tanks. Taking the thermal stratification, pressure rise phenomenon and natural convection of liquid helium storage tank as research objects, based on the Lee model, the finite volume method was used for the first time to study the non-steady-state thermal response which simultaneously considering the change of liquid helium and vapor helium properties. Thus the tank pressure, temperature and flow rate distribution at different times were obtained, and the effects of evaporation rate and filling rate on the pressure rise and temperature rise of the storage tank were analyzed. The research shows that with the increase of simulation time, the liquid helium shows thermal stratification. In addition, the pressure is distributed in a stepwise manner which is lower in the upper part, and higher in the lower part, while the pressure distribution in the gas phase space is more uniform; After the liquid helium in the near-wall area is heated, it rises along the wall surface to the free liquid surface under buoyancy lift, and then flows into the main flow area of the liquid helium; As the evaporation rate increases, the temperature rise and pressure rise rate in the gas phase space of the storage tank increase, while the filling rate have less influence on the temperature rise and pressure rise rate.This work provides guidance for non-destructive storage and transportation theory of liquid helium storage tanks.
Free
COVID-19 Patient Health Monitoring System
Scientific article
The system proposed can be used to regular checkup of the COVID patients while maintaining the social distancing. Also, the data sensed by the sensors is directly sent to doctor, reducing the cost of paying regular visits to doctor. The Iot platform used in the system helps to transfer the real time patient’s data remotely to host device. Daily health record can be maintained and can be viewed easily on graphs charts ease for doctors to see any abrupt changes in oxygen level or rise in temperature. To track the patient health micro-controller is in turn interfaced to an LCD display and wi-fi connection to send the data to the web-server (wireless sensing node). In case of any abrupt changes in patient heart-rate or body temperature alert is sent about the patient using IoT. This system also shows patients temperature and heartbeat tracked live data with timestamps over the Internetwork.
Free
Scientific article
In most studies applying deep learning to medical image analysis, classification accuracy is the sole optimization criterion, while the computational cost of the resulting models remains undocumented: parameter and operation counts are reported only occasionally, and inference latency on a central processing unit is almost never published. This makes an informed model choice impossible for institutions without graphics accelerators, such as district hospitals, mobile diagnostic units and field hospitals. This paper experimentally investigates the trade-off between diagnostic performance and computational resources for four-class chest X-ray classification. The computational cost of 15 widely used architectures was evaluated with a single tool at an input resolution of 224×224, and the gap between the heaviest and the lightest architecture reaches a factor of 323. Four models – a heavy baseline, two lightweight architectures, and a custom compact network – were trained under a unified protocol on the COVID-19 Radiography Database (21,165 images across four classes) and profiled strictly on a CPU with a batch size of one. The three standard architectures were initialized with ImageNet-pretrained weights, and the custom network was trained from scratch. Each architecture was trained across three independent random seeds on a fixed data split, so all performance metrics are reported as mean ± std, and every comparison is accompanied by a confidence interval and a significance test. The lightweight architecture trailed the baseline by only 1.24 percentage points in macro-F1 (95% CI: 0.84 to 1.63, p = 0.005) while requiring 69 times fewer operations, achieving 19.3 times lower latency, and reducing model size by a factor of 15.4; the efficiency metric, defined as macro-F1 per GFLOP, differs by a factor of 68. Additionally, post-training quantization systematically failed across all three runs for both architectures combining depthwise convolutions with Squeeze-and-Excitation blocks, whereas the same pipeline quantized the remaining models without statistically significant loss. Neither depthwise convolutions nor the Hard-Swish activation accounted for the failure; instead, the number of channel-wise multiplication nodes in the exported graph separated the two groups exactly, making the risk identifiable directly from the static graph prior to training. The practical implication is that the quantizability of a lightweight model must not be assumed; it must be verified empirically.
Free
Calculation of Failure Probability of Series and Parallel Systems for Imprecise Probability
Scientific article
In the situation that unit failure probability is imprecise when calculation the failure probability of system, classical probability method is not applicable, and the analysis result of interval method is coarse. To calculate the reliability of series and parallel systems in above situation, D-S evidence theory was used to represent the unit failure probability. Multi-sources information was fused, and belief and plausibility function were used to calculate the reliability of series and parallel systems by evidential reasoning. By this mean, lower and upper bounds of probability distribution of system failure probability were obtained. Simulation result shows that the proposed method is preferable to deal with the imprecise probability in reliability calculation, and can get additional information when compare with interval analysis method.
Free
Characterization and Anti-proliferative Activity of Ethanol Extract of Tartary Buckwheat
Scientific article
Tartary buckwheat (Fagopyrum tataricum) has been consumed in China as health food and medicine. In this study, the ethanol extract (EE) of tartary buckwheat was prepared. The major flavonoid in EE was identified as rutin by HPLC-PDA-ESI/MS and the content of rutin in EE was determined as 27.4 % by HPLC. In the anticancer assay, it was found EE could inhibit the proliferation of the human hepatocellular liver carcinoma Hep G2 cell line in a dose-dependent manner. The increase of an early apoptotic population was observed by both annexin-FITC and PI staining, which suggested that EE could induce Hep G2 cells to enter into apoptosis.
Free
Classification of small sets of images with pre-trained neural networks
Scientific article
Nowadays the rise of the artificial intelligence is with high speed. Even we are far away from the moment when machines are going to make decisions instead of human beings, the development in some fields of artificial intelligence is astonishing. Deep neural networks are such a filed. They are in a big expansion in a new millennium. Their application is wide: they are used in processing images, video, speech, audio, and text. In the last decade, researches put special attention and resources in the development of special kind of neural networks, convolutional neural networks. These networks have been widely applied to a variety of pattern recognition problems. Convolutional neural networks were trained on millions of images and it is difficult to outperform the accuracies that have been achieved. On the other hand, when we have a small dataset to train the network, there is no success to do it from a scratch. This article exploits the technique of transfer learning for classifying the images of small datasets. It consists fine-tuning of the pre-trained neural network. Here in details is presented the selection of hyper parameters in such networks, in order to maximize the classification accuracy. In the end, the directions have been proposed for the selection of the hyper parameters and of the pre-trained network which can be suitable for transfer learning.
Free
Color Image Segmentation Using Level Set Method With Initialization Mask in Multiple Color Spaces
Scientific article
The aim of image segmentation in imaging science is to solve the problem of partitioning an image into smaller disjoint homogeneous regions that share similar attributes. The improvement of level set method (LSM) based on Chan-Vese (C-V) model with initialization mask for vector image segmentation in multiple color spaces is studied here. And simultaneously, the final segmentation is completed by a simple labeling scheme. Then the comparative study of the refined C-V model is done in multiple color spaces. The experimental results illustrate that the optimized C-V model leads faster and better segmentation results with robustness to noise and good adaptability in RGB, CIE XYZ, and YCbCr color spaces where the results of test image changes little. But it has made mistakes in HSV and CIE L*a*b* color model. Moreover, these color spaces, i.e. h1h2h3, produce poor segmentation on the reliability and accuracy of a set of test images by performance analysis with evaluation indicators.
Free
Colour, Texture, and Shape Features based Object Recognition Using Distance Measures
Scientific article
Object recognition is the recognizing process of objects into semantically expressive classes using its visual insides. Classification of objects becomes complex and challenging task because of its size, poor image quality, occlusion, scaling, geometric distortion, lightening, etc. In this paper, global descriptors that means Color, Texture, and Shape features are used to recognize object. Color histogram is used to obtain the color content, texture content is obtained using Gabor wavelet, and shape content is extracted using Hough transform. These low level or global features are used for creating feature vector. Distance measure is used to find the 1-Nearest Neighbor from the training images i.e. object with minimum distance or maximum similarity with visual contents of the query image. The class of that training image is the predicted label of the query image. We have used twelve different distance measures: some are metrics, some are non-metrics and finally, their recognition accuracy is compared. Ensemble of these distance measures is also used for object recognition in the image. We evaluate this method on a publicly available object-recognition dataset: Columbia Object Image Library (COIL-100) dataset. The experiments show that the recognized results outperform many state-of-the-art methods.
Free
Combining Local Binary Patterns and Visual Attention for Face Recognition
Scientific article
Effectiveness of local binary pattern (LBP) for face recognition has been proven. But the weight of weighted LBP is difficult to determine. In this paper, we proposed a biologically plausible approach to set the weight automatically. Combining LBP and visual attention, a weight map can be constructed by summing over the saliency map. The weight map outlines salient information in the image and helpful for recognition. Experimental results show that the presented method is efficient and effective.
Free
Communication Challenges for Distributed Teams
Scientific article
The paper discusses the issues and problems faced by individuals working in diverse or distributed teams especially in the software industry. The research shows a feedback and trend of the teams that work in distributed environments letting us understand the behaviors and figuring out solutions for the problems of communication. The trend was already set in a few organizations that provided outsourcing services to their clients but after the recent issue of the pandemic almost all organizations switched towards this trend and started focusing on working in distributed teams and not on site. This ensured the safety of the employees as well as it let them keep their work and performance undisturbed. But in the practice of working in distributed teams the major factor that affected the performance was the problem in communication, either due to the available mediums (online meeting software products) or the internet connections. The only factor affecting the performances somehow was the communication problem factor and we have addressed this in our paper, moreover a survey is also done through a questionnaire with the people of software industries in Pakistan and a result is generated for how the people feel working in distributed environments and what hurdles they faced in communication.
Free
Scientific article
It has analyzed the transition processes for the content of the revenue equivalence principle. It used the revenue equivalence principle to derive the equilibrium bidding strategies in the first, second and third price auction and the all-pay auction from the expected payment of the bidders. Based on this, it compared the equilibrium bidding strategies in the four auction forms and got some useful conclusions.
Free
Comparative Analysis of Various SRAM Cells with Low Power, High Read Stability and Low Area
Scientific article
SRAMs are very important part of today's movable devices like laptops and mobile phones. Different SRAM cells of different number of transistors have their own respective advantages and drawbacks. In this work an attempt has been made to reduce the leakage power by adding some transistors. Each SRAM cell provides an efficient way to reduce the leakage power, but disadvantage of each SRAM cell limit the application of them. In this paper, the study and transient analysis on four different SRAM cells has been carried out and compared with respect to various parameters like power dissipation, delay and area. The simulation is carried out in 180nm CMOS technology using tanner tools. Layout is carried out using microwind.
Free
Scientific article
In this study, the hyperparameters of Stochastic Gradient Boosting, LightGBM, and Histogram-based Gradient Boosting models were optimized using evolutionary algorithms. The main goal of this research is to find the most effective combination of hyperparameters for each model. Their goal is to increase accuracy and improve computational efficiency. The study was conducted on the basis of real clinical data. The data were obtained from the Samarkand City Endocrinology Center, using a dataset of diabetes-related and clinical indicators. The data were initially cleaned and processed using normalization, imputation, and 5-point cross-validation. We used five evolutionary strategies to fine-tune the main hyperparameters: Differential Evolution (DE), Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Covariance Matrix Adaptation Evolution Strategy (CMA-ES), and Dynamic Ensemble with Hyperband (DEHB).In the optimization, the F1-macro was chosen as the main fitness function. This allows for a balanced accuracy assessment for multi-class classification. Across all three prototypes, evolutionary tuning consistently led to better results than Grid Search and Random Search. SGB+DE accuracy 0.9098, F1-macro 0.9096 gave the best results for the single model, whereas DEHB and CMA-ES worked better for LightGBM and HGB, respectively. The absolute increases above the unadjusted baselines ranged from 0.3 to 0.6 percentage points. They were small but could be repeated reliably across all evaluation criteria. During this study, the models were evaluated on the metrics of accuracy, precision, recall, and F1-macro.
Free
Scientific article
Early and accurate detection of citrus diseases is essential for maintaining fruit quality, minimising crop losses, and supporting sustainable agricultural production. Automated image-based diagnostic systems offer a scalable alternative to conventional manual inspection, which is often time-intensive, subjective, and susceptible to diagnostic variability. This study presents a transfer-learning-based deep learning framework for automated classification of citrus fruit diseases using a curated image collection derived from Food and Agriculture Organisation (FAO) resources. The dataset was organised into two classification groups: lemon and orange. The lemon dataset initially comprised 208 images across four classes: Healthy, Canker, Mold, and Scab, while the orange dataset contained 2,240 images across two classes. To address data scarcity and improve model generalisation, targeted data augmentation involving zooming (0.8–1.2), rotation (35°), and horizontal flipping was employed, increasing the datasets to 1,600 and 4,000 images for lemon and orange, respectively. Five ImageNet-pretrained convolutional neural network architectures—VGG16, ResNet50, InceptionV3, DenseNet121, and EfficientNetB0—were fine-tuned and evaluated using a stratified 70:30 training–testing protocol with five-fold cross-validation. Model performance was assessed using accuracy, precision, recall, and F1-score under a standardized experimental configuration. The results demonstrate that InceptionV3 achieved the highest classification accuracy of 90.0% on the four-class lemon dataset, while DenseNet121 obtained the best accuracy of 93.0% on the binary orange dataset. These findings indicate that appropriate transfer learning and targeted augmentation can substantially improve classification performance and model generalisation, particularly in limited-data agricultural imaging scenarios. The proposed benchmarking framework enables systematic, controlled comparisons of fine-tuned CNN architectures under consistent preprocessing, augmentation, training, and evaluation conditions. The resulting models can be integrated into real-time citrus disease diagnostic systems and deployed on resource-constrained platforms, including mobile and edge-computing devices, supporting scalable precision-agriculture applications.
Free
Comparative Evaluation of Supervised Learning Algorithms for Breast Cancer Classification
Scientific article
Breast cancer is a leading cause of mortality among women worldwide, particularly in developing countries. Accurate and early diagnosis is crucial to improve patient outcomes. This study compares the performance of three supervised machine learning algorithms—Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Random Forest (RF)—in classifying breast cancer cases using the Breast Cancer Wisconsin dataset. The dataset consists of 569 instances with 33 features, categorized into malignant and benign classes. Each method was evaluated based on its classification accuracy. The results show that Random Forest achieved the highest accuracy at 94.07%, outperforming SVM with 90.06% and KNN with 90.00%. The findings suggest that Random Forest provides the most reliable performance for breast cancer classification within the scope of this dataset. This study highlights the importance of selecting an appropriate algorithm to enhance diagnostic precision and recommends Random Forest as an effective method for similar classification tasks.
Free
Comparative Exploration of the Contribution of Reinforcement Learning in Robotic Surgery
Scientific article
Problem: The precision, efficiency, and safety of surgical procedures need significant improvements. Traditional methods are limited by human capabilities, and existing robotic systems lack the advanced adaptability required for complex surgical tasks. The integration of reinforcement learning (RL) into robotic surgery represents a potential revolution in the medical field. Methods: This comparative review synthesizes recent progress in RL applications for robotic surgery. It highlights innovative methodologies and successful applications of RL, focusing on advanced simulations to train RL agents and the importance of human demonstrations in the learning process. Results: Emerging trends such as the effective use of simulations and human demonstrations to enhance RL in robotic surgery are identified. The review also discusses challenges associated with RL applications, emphasizing the need for clinical validation and ensuring patient safety. Conclusion: The transformative potential of RL in robotic surgery is evident, though challenges remain. Future work should prioritize clinical validation, patient safety, and interdisciplinary collaboration.
Free
Comparative Study of Website Sitemap Feature as Design Issue in Various Websites
Scientific article
Websites are very important for communicating organizational aims and goals in this current era of information technology. Different institutions / organizations put lots of efforts to portray complete information on beautifully designed websites. Lot of efforts are required by the designer / concerned organization to design websites according to the different website design standards to provide users with all the facilities of the concerned institutions / organizations online, which act as an online agent through which a user can get his work done without physically visiting the organizations. With this the responsibility of the designer and the concerned institutions / organizations increases manifold so that the websites behavior should remain interactive and quick enough for the user to avail provided facilities of the website comfortably. Sitemap provides with a logical overview of the path through which a user can navigate through a website. Without efficient and user-friendly navigation, the user is likely to get confused, lost, or frustrated and leave the site without getting their work done. Sitemap gives users with the graphical representation of Web pages on a Web Site and also present them with a complete alternative method of navigating the site. Author in this paper developed an online tool using .NET Framework using C# to study webpage sitemap feature as Design issue in various categories of the websites like Government, Commercial, Educational, Social networking and Job portals. The automated tool developed by author function on the basis of the different standards prescribed in W3C WCAG 1.0 guidelines in priority 2 checkpoints 13.3 [5] . the developed tool act like a parser and renders the complete code of the website and produces results by determining the presence of the sitemap facility within the websites for easy content retrieval. The results produced shows that out of the five different categories of websites employed for analysis the commercial websites undertaken follow the maximum standards as far as sitemap parameter is concerned.
Free
