Search results

The purpose of this work is to present an approach for autonomous detection of eye disease in fundus images. Furthermore, this work presents an improved variant of the Tiny YOLOv7 model developed specifically for eye disease detection. The model proposed in this work is a highly useful tool for the development of applications for autonomous detection of eye diseases in fundus images that can help and assist ophthalmologists.

The approach adopted to carry out this work is twofold. Firstly, a richly annotated dataset consisting of eye disease classes, namely, cataract, glaucoma, retinal disease and normal eye, was created. Secondly, an improved variant of the Tiny YOLOv7 model was developed and proposed as EYE-YOLO. The proposed EYE-YOLO model has been developed by integrating multi-spatial pyramid pooling in the feature extraction network and Focal-EIOU loss in the detection network of the Tiny YOLOv7 model. Moreover, at run time, the mosaic augmentation strategy has been utilized with the proposed model to achieve benchmark results. Further, evaluations have been carried out for performance metrics, namely, precision, recall, F1 Score, average precision (AP) and mean average precision (mAP).

The proposed EYE-YOLO achieved 28% higher precision, 18% higher recall, 24% higher F1 Score and 30.81% higher mAP than the Tiny YOLOv7 model. Moreover, in terms of AP for each class of the employed dataset, it achieved 9.74% higher AP for cataract, 27.73% higher AP for glaucoma, 72.50% higher AP for retina disease and 13.26% higher AP for normal eye. In comparison to the state-of-the-art Tiny YOLOv5, Tiny YOLOv6 and Tiny YOLOv8 models, the proposed EYE-YOLO achieved 6–23.32% higher mAP.

This work addresses the problem of eye disease recognition as a bounding box regression and detection problem. Whereas, the work in the related research is largely based on eye disease classification. The other highlight of this work is to propose a richly annotated dataset for different eye diseases useful for training deep learning-based object detectors. The major highlight of this work lies in the proposal of an improved variant of the Tiny YOLOv7 model focusing on eye disease detection. The proposed modifications in the Tiny YOLOv7 aided the proposed model in achieving better results as compared to the state-of-the-art Tiny YOLOv8 and YOLOv8 Nano.

The purpose of this paper is to describe a new approach to sentence representation learning leading to text classification using Bidirectional Encoder Representations from Transformers (BERT) embeddings. This work proposes a novel BERT-convolutional neural network (CNN)-based model for sentence representation learning and text classification. The proposed model can be used by industries that work in the area of classification of similarity scores between the texts and sentiments and opinion analysis.

The approach developed is based on the use of the BERT model to provide distinct features from its transformer encoder layers to the CNNs to achieve multi-layer feature fusion. To achieve multi-layer feature fusion, the distinct feature vectors of the last three layers of the BERT are passed to three separate CNN layers to generate a rich feature representation that can be used for extracting the keywords in the sentences. For sentence representation learning and text classification, the proposed model is trained and tested on the Stanford Sentiment Treebank-2 (SST-2) data set for sentiment analysis and the Quora Question Pair (QQP) data set for sentence classification. To obtain benchmark results, a selective training approach has been applied with the proposed model.

On the SST-2 data set, the proposed model achieved an accuracy of 92.90%, whereas, on the QQP data set, it achieved an accuracy of 91.51%. For other evaluation metrics such as precision, recall and F1 Score, the results obtained are overwhelming. The results with the proposed model are 1.17%–1.2% better as compared to the original BERT model on the SST-2 and QQP data sets.

The novelty of the proposed model lies in the multi-layer feature fusion between the last three layers of the BERT model with CNN layers and the selective training approach based on gated pruning to achieve benchmark results.

This work aims to present a deep learning model for face mask detection in surveillance environments such as automatic teller machines (ATMs), banks, etc. to identify persons wearing face masks. In surveillance environments, complete visibility of the face area is a guideline, and criminals and law offenders commit crimes by hiding their faces behind a face mask. The face mask detector model proposed in this work can be used as a tool and integrated with surveillance cameras in autonomous surveillance environments to identify and catch law offenders and criminals.

The proposed face mask detector is developed by integrating the residual network (ResNet)34 feature extractor on top of three You Only Look Once (YOLO) detection layers along with the usage of the spatial pyramid pooling (SPP) layer to extract a rich and dense feature map. Furthermore, at the training time, data augmentation operations such as Mosaic and MixUp have been applied to the feature extraction network so that it can get trained with images of varying complexities. The proposed detector is trained and tested over a custom face mask detection dataset consisting of 52,635 images. For validation, comparisons have been provided with the performance of YOLO v1, v2, tiny YOLO v1, v2, v3 and v4 and other benchmark work present in the literature by evaluating performance metrics such as precision, recall, F1 score, mean average precision (mAP) for the overall dataset and average precision (AP) for each class of the dataset.

The proposed face mask detector achieved 4.75–9.75 per cent higher detection accuracy in terms of mAP, 5–31 per cent higher AP for detection of faces with masks and, specifically, 2–30 per cent higher AP for detection of face masks on the face region as compared to the tested baseline variants of YOLO. Furthermore, the usage of the ResNet34 feature extractor and SPP layer in the proposed detection model reduced the training time and the detection time. The proposed face mask detection model can perform detection over an image in 0.45 s, which is 0.2–0.15 s lesser than that for other tested YOLO variants, thus making the proposed detection model perform detections at a higher speed.

The proposed face mask detector model can be utilized as a tool to detect persons with face masks who are a potential threat to the automatic surveillance environments such as ATMs, banks, airport security checks, etc. The other research implication of the proposed work is that it can be trained and tested for other object detection problems such as cancer detection in images, fish species detection, vehicle detection, etc.

The proposed face mask detector can be integrated with automatic surveillance systems and used as a tool to detect persons with face masks who are potential threats to ATMs, banks, etc. and in the present times of COVID-19 to detect if the people are following a COVID-appropriate behavior of wearing a face mask or not in the public areas.

The novelty of this work lies in the usage of the ResNet34 feature extractor with YOLO detection layers, which makes the proposed model a compact and powerful convolutional neural-network-based face mask detector model. Furthermore, the SPP layer has been applied to the ResNet34 feature extractor to make it able to extract a rich and dense feature map. The other novelty of the present work is the implementation of Mosaic and MixUp data augmentation in the training network that provided the feature extractor with 3× images of varying complexities and orientations and further aided in achieving higher detection accuracy. The proposed model is novel in terms of extracting rich features, performing augmentation at the training time and achieving high detection accuracy while maintaining the detection speed.

This paper aims to perform an experimental investigation on the impact strength, compressive strength, tensile strength and flexural strength of fly ash-based green composites and to compare with these polyvinyl chloride (PVC), high density polyethylene (HDPE) and low density polyethylene (LDPE).

Fly ash-based polymer matrix composites (FA-PMCs) were fabricated using hand layup method. Composites containing 100 g by weight fly ash particles, 100 g by weight brick dust particles and 50 g by weight chopped glass fiber particles were processed. Impact strength, compressive strength, tensile strength and flexural strength of composites have been measured and compared with PVC, HDPE and LDPE. Impact strength of the FA-PMC is higher than that of PVC, HDPE and LDPE. Structural analysis of pipes, gears and axial flow blade was verified using ANSYS. Barlou’s condition for pipes, Lewis–Buckingham approach for gears and case-based analysis for axial flow blades were carried out and verified.

Pipes, gears and axial flow blades made form fly ash-based composites were found to exhibit improved thermal resistance (i.e. better temperature independence for mechanical operations), higher impact strength and longer life compared to those made from PVC, HDPE and LDPE. Moreover, the eco-friendly nature of the raw materials used for fabricating the composite brings into its quiver a new dimension of appeal.

Experimental investigation on the impact strength, compressive strength, tensile strength and flexural strength of fly ash-based green composites has not been attempted yet.

Positive experiences with working from home (WFH) during the Corona pandemic (COVID-19) have motivated many employees to continue WFH after the pandemic. However, factors influencing employees' WFH intentions against the backdrop of experiences during pandemic-induced enforced working from home (EWFH) are heterogeneous. This study investigates factors linked to information technology (IT) professionals' WFH intentions.

This mixed-methods study with 92 IT professionals examines the effects of seven predictors for IT professionals' WFH intentions. The predictors are categorized according to the trichotomy of (1) characteristics of the worker, (2) characteristics of the workspace and (3) the work context. Structural equation modeling is used to analyze the quantitative survey data. In addition, IT professionals' responses to six open questions in which they reflect on past experiences and envision future work are examined.

Quantitative results suggest that characteristics of the worker, such as segmentation preference, are influencing WFH intentions stronger than characteristics of the workspace or the work context. Furthermore, perceived productivity during EWFH and gender significantly predict WFH intentions. Contextualizing these quantitative insights, the qualitative data provides a rich yet heterogeneous list of factors why IT professionals prefer (not) to work from home.

Reasons influencing WFH intentions vary due to individual preferences and constraints. Therefore, a differentiated organizational approach is recommended for designing future work arrangements. In addition, the findings suggest that team contracts to formalize working patterns, e.g. to agree on the needed number of physical meetings, can be helpful levers to reduce the complexity of future work that is most likely a mix of WFH and office arrangements.

This study extends literature reflecting on COVID-19-induced changes, specifically the emerging debate about why employees want to continue WFH. It is crucial for researchers and practitioners to understand which factors influence IT professionals' WFH intentions and how they impact the design and implementation of future hybrid work arrangements.

Even the greatest developed countries have capitulated to the destructions imposed on the global supply systems, as the COVID-19 pandemic has revealed. The purpose of this study is to explore human resource sourcing strategies for managing supply chain performance during the COVID-19 outbreak. There are six human resource sourcing strategies such as outsourcing, near sourcing, integration, the requirement of suppliers, joint ventures and virtual enterprise that are considered to measure supply chain performance.

Based on collecting data from the potential respondents of Indian manufacturing companies, the elevation of human resource sourcing strategies to supply chain performance is measured considering the multiple regression analysis techniques.

The results of the study revealed that four of the six hypotheses have a significant and positive relationship with supply chain performance during the COVID-19 outbreak while two hypotheses are partially supported that lent good support to this study.

In this critical situation, this study will enable managers and practitioners to support the business in giving customers the best services on time.

The novelty of this study is to identify the key human resource sourcing strategies by using multiple regression analysis methods, considering the case of Indian manufacturing companies to measure their supply chain performance during the COVID-19 outbreak era.

This study aims to examine the implementation of agile practices in Industry 4.0 to assess the financial performance measurements of manufacturing firms. It also investigates the relationship between supply chain performance and financial performance.

The study is based on an experimental research design by collecting data from 329 responses from key officials of manufacturing firms. The analyses are carried out to explore this modern concept with the help of the SPSS program, which is used to conduct a confirmatory factor and reliability analysis and Smart-partial least square (PLS) version 4.0 with structural equation modeling.

This research demonstrates the positive effect agile supply chain strategies in Industry 4.0 may have on manufacturing companies' financial performance as a whole. Everything throughout the supply chain in Industry 4.0, from the manufacturers to the end users, is taken into account as a potential performance booster. The values obtained from the model's study show that it is both dependable and effective, surpassing the threshold for such claims. The research is supported by factors like customer involvement (CUS), continuous improvement (CI), integration (INT), modularity (MOD), management style (MS) and supplier involvement (SI) but is undermined by factors including postponement (PPT).

According to the findings of the study, Industry 4.0 firms' financial performance and overall competitiveness are significantly improved when their supply chains are more agile. A more agile supply chain helps businesses to more rapidly adapt to shifts in consumer demand, shorten the amount of time it takes to produce a product, enhance product quality and boost customer happiness. As a consequence of this, there will be an increase in revenue, an improvement in profitability and continued sustainable growth.

There are literary works available on agile practices in various fields, but the current study outlines the need to understand how supply chains perform financially under the mediating effect of agile supply chains in Industry 4.0 which contribute most to the organization's success. The study will aid companies in understanding how agile practices will further the overall performance of the organization financially.

AA2014 is a copper-based alloy and is typically used for production of complex machined components, given its better machinability. The purpose of this paper was to study the effects of variation in weight percentage of ceramic Al₂O₃ particulates during electrical discharge machining (EDM) of stir cast AA2014 composites. Scanning electron microscopy (SEM) examination was carried out to study characteristics of EDMed surface of Al₂O₃/AA2014 composites.

The effect of machining parameters on performance measures during sinker EDM of stir cast Al₂O₃/AA2014 composites was examined by “one factor at a time” (OFAT) method. The stir cast samples were obtained by using three levels of weight percentage of Al₂O₃ particulates, i.e. 0 Wt.%, 10 Wt.% and 20 Wt.% with density 1.87 g/cc, 2.35 g/cc and 2.98 g/cc respectively. Machining parameters varied were peak current (1-30 amp), discharge voltage (30-100 V), pulse on time (15-300 µs) and pulse off time (15-450 µs) to study their influence on material removal rate (MRR), tool wear rate (TWR) and surface roughness (SR).

MRR and SR decreased with an increase in weight percentage of ceramic Al₂O₃ particulates at the expense of TWR. This was attributed to increased microhardness for reinforced stir cast composites. However, microhardness of EDMed samples at fixed values of machining parameters, i.e. 9 amp current, 60 V voltage, 90 µs pulse off time and 90 µs pulse on time reduced by 58.34, 52.25 and 46.85 per cent for stir cast AA2014, 10 Wt.% Al₂O₃/AA2014 and 20 Wt.% Al₂O₃/AA2014, respectively. SEM and quantitative energy dispersive spectroscopy (EDS) analysis revealed ceramic Al₂O₃ particulate thermal spalling in 20 Wt.% Al₂O₃/AA2014 composite. This was because of increased particulate weight percentage leading to steep temperature gradients in between layers of base material and heat affected zone.

This work was an essential step to assess the machinability for material design of Al₂O₃ reinforced aluminium metal matrix composites (AMMCs). Experimental investigation on sinker EDM of high weight fraction of particulates in AA2014, i.e. 10 Wt.% Al₂O₃ and 20 Wt.% Al₂O₃, has not been reported in archival literature. The AMMCs were EDMed at variable peak currents, voltages, pulse on and pulse off times. The effects of process parameters on MRR, TWR and SR were analysed with comparisons made to show the effect of Al₂O₃ particulate contents.

The purpose of this study is to investigate the challenges before India for electric vehicle (EV) adoption by 2030. The study further looks into the measures taken by the Government of India (GOI) to promote research and development in EV sector and what is yet to be done.

In the present study, the challenges are identified allied to the commercialization of EVs in India. The data are collected, analyzed and compiled through secondary sources. The secondary data give a concise insight and comprehensive information regarding what is occurring around the globe as well as in the Indian context. Further, the challenges are investigated through a focus group study consisting of 11 participants from industry and academia.

The findings from the study are the critical roles of sharing economy and public utilities in the promotion of EV adoption, given the high cost of EV, lack of infrastructure and poor purchasing power of Indian customers. The sharing economy perspective provides various opportunities for the government to manage the resources (electric-powered transport system) optimally. Further, the study compares the global perspective in assigning the target figures.

The study highlights the facilitating role of the shared format in EV technology promotion but ignores the hurdles that can come in its implementations. Also, the focus group study has its limitation as it relies more on participants' perceptions and opinions.

The present study assists GOI and various stakeholders in having a realistic plan rather than daydreaming with overambitious goals. The diffusion of technology as a shared format (especially in the context of EV) has not been academically approached in the past literature.

Employee engagement (EE) continues to be one of the most difficult challenges for organizations today. Numerous factors have been linked to EE, according to studies. However, the necessary human resource management (HRM) strategies and systems for enhancing EE have not yet been developed. It is questionable if all employees inside the company require the same HRM strategies, to boost engagement as one size does not fit all. Therefore, it is necessary to create employee profiles based on factors associated with EE. This study aims to develop employee profiles based on engagement dimensions and outcomes. It seeks to comprehend the relationship between engagement level and factors such as age, years of service and employment grade.

Using latent profile analysis (LPA), we identified five EE profiles (highly engaged, engaged, moderately engaged, disengaged and highly disengaged). These five profiles were characterized by five EE dimensions (Culture Dimensions, Leadership Dimensions, People Process, Business alignment Dimension and Job Dimension) and EE outcomes (Say, Stay and Strive).

The study revealed that Engaged profiles exhibited low stay outcomes. The highest percentage of disengaged employees fall under 25 years of age with less than 5 years of experience and are at the entry level.

The study highlights the significance of the people processes dimensions in enhancing engagement. Profiles with low people process dimensions showed high disengagement. Person-centered LPA adds and complements variable-centered approach to develop a better understanding of EE and help organizations devise more personalized strategies. The study would be of interest to both academics and practitioners.

The novelty of this study lies in its attempt to model the employee profiles to comprehend the relationship between engagement levels using LPA.