Comparison of Sequential and Nonsequential Models for Spanish to English Machine Translation

Deep Gandhi, Jash Mehta and Pranit Bari, Department of Computer Engineering, Dwakradas J Sanghvi College of Engineering, India

ABSTRACT

Machine Translation has been at the forefront as one of the big applications of Natural Language Processing. The task of translating Spanish to English is the one to be dealt with for this paper. We propose and compare various sequential and non-sequential models in order to determine the best one for the task. All of the models compared have an encoder-decoder architecture. These predictions are also then used to compare various prediction algorithms in order to reconstruct the translations by preserving the semantic meaning from one language to the other. Thus, for this task, various prediction methods such as Top-k sampling and Nucleus Sampling are to be used and determined which is the best instead of just using the Greedy approach as context is of importance for reconstruction here. Our final recommendation achieves an accuracy of 52.9 and BLEU score 33.23 on the UN Languages dataset.

KEYWORDS

Natural Language Processing, Neural Machine Translation, Deep Learning, Neural Networks, Seq2Seq.

End-to-End Spanish-English Sequence Learning Translation Model

Vidhu Mitha Goutham1 and Ruma Mukherjee, Unisys Corporation, Bangalore, India

ABSTRACT

The low availability of unlimited, dynamic-access models for specific languages, makes it hard for corporate users to adopt quick translation techniques and incorporate them into product solutions. As translation tasks increasingly require a dynamic sequence-learning curve, cost-free opensource models are scarce. We survey and compare current translation techniques and propose a modified sequence-tosequence model repurposed with attention techniques. Sequence learning using an encoder-decoder model is paving the path for higher precision levels in translation. Using a CNN encoder and RNN decoder background, we use Fairseq tools to produce an end-to-end bilingually trained Spanish-English machine translation model including source language detection. We acquire competitive results using a duolingocorpus-trained model to provide ready-made plug-in use for compound sentences and document translations. Our model serves a robust system for large, organizational data-translation needs. While acknowledging its shortcomings and future scope- it also identifies itself as a well-optimized deep neural network model and solution.

KEYWORDS

Translation, Spanish, Fairseq, sequence-to-sequence, encoder-decoder.

Foreign Language Education and Programmatic Teaching

Saide Sadikoglu, Faculty of Tourism, Near East University, Nicosia, North Cyprus, Mersin 10 Turkey

ABSTRACT

In the modern globalized world, commercial, economic, social, and cultural relations require the use of language for international communication. A literature review shows thatforeign language education programs are not adequately prepared. The problem of foreign language education still exists to a large extent todaythe most important reason for this is that existing teaching approaches are used in an incomplete or unscheduled manner. This studyaims to examine the foreign language education programs in secondary education and higher education in terms of theoretical aspects, teaching methods, target language level as well as measurement and evaluation.

KEYWORDS

Education Program, Foreign Language Teaching, Technology approach.

Statistical Language Models of English: Analysis of Dictionary Properties

Anastasia Malashina, HSE University, Computer Security Department, Moscow, Russia

ABSTRACT

This article examines two language models of the English language when considering different body sizes and their statistical analysis. Using a series of experiments, we determine some informational properties of the language, such as the vocabulary size, the corpus coverage, and the n-gram entropy. The empirical results obtained have a minimal discrepancy between the calculated and theoretical characteristics due to the large volume of the corpus. Graphical extrapolation allows us to predict further values as the corpus size increases. We propose an approach to the calculation of theoretical corpus coverage. The results show the crucial role of corpus size to increase the vocabulary coverage. In addition, we discuss an effective methodology for determining entropy values for short texts based on dictionary volumes. The results show that the entropy values of n-grams are close to the real values for the English language.

KEYWORDS

n-gram dictionaries, n-gram entropy, meaningful texts, Zip’s law, lemmatization, text coverage.

IoT-BDMS : Securing IoT Devices with Hyperledger Fabric Blockchain

Nathalie BANOUN, Nafissatou DIARRA, Department of Research and Innovation, Capgemini, France

ABSTRACT

IoT is a rapidly evolving field with an increasing number of connected devices. This naturally leads to a need to ensure good scalability but also to guarantee the identity of devices, for better security. Most of existing solutions for identifying IoT devices are centralized (CA server), which results in lower fault tolerance and in less scalability. To address these issues, we introduce in this paper a new IoT Device Management System based on the Blockchain technology (IoT-BDMS). Our system offers two services through two Smart Contracts deployed on a multi-channel Hyperledger Fabric network: an Identification Smart Contract (ISC) to manage the devices identities stored over multiple channels, and an Authentication Smart Contract (ASC) to validate authentication requests from devices. An identity is generated by involving the actors of the device’s ecosystem and evolves according to the device’s lifecycle.

KEYWORDS

Blockchain, IoT Security, Hyperledger Fabric, Smart Contracts, Authentication.

Blockchain-Based Approach to Foster Student Engagement on Campus

Ritu Gala, Eshita Shukla, Nidhee Kamble and Revathi Vijayaraghavan, Department of Computer Engineering and Information Technology, Veermata Jijabai Technological Institute, Mumbai, India

ABSTRACT

On-campus activities like positions of responsibility in campus amenities and participation in research, benefit the students as well as the university, while also making students financially self-sufficient to a certain extent. However, this student participation is stymied by lack awareness, and motivation. Significant impetus to innovation and student participation can be provided by incentivisation of these activities. In this paper, we propose a system to create a blockchain-based economy, to incentivise students with empirical benefits or monetary awards calculated using objective algorithms. The incentivisation algorithms have been designed for three promising use cases: research work, positions of responsibility in universities, and crowdfunding. The demonstrated implementation of this system utilises VJTI Chain, an already established Proof of Authority blockchain in VJTI Mumbai, India, so that its existing infrastructure finds a more extensive usage in the cutting-edge technology ecosystem. It uses VJCoins for awards, a currency native to the blockchain’s circular economy within the campus.

KEYWORDS

Blockchain, Distributed Ledger Technology (DLT), Learning incentivization.

A Machine Learning Approach to Evaluating student’s Learning Affects: An Application to on-line Learning platform

Asaju Christine and Hima Vadapalli, Department of Computer Science, University of the Witwatersrand, Johannesburg, South Africa

ABSTRACT

A major obstacle experienced during online classes is that, there is no proper evaluation and feed back to tutors unlike in the conventional classroom teaching and learning process, where teachers are able to evaluate the students through their facial emotions expression, body languages, gesture, etc. This study, therefore, considers a machine learning approach to evaluate and provide feed back to teachers in an online class, using facial emotion recognition, mapping of emotions to various learning affects, and using it as feed back to teachers. The study adopts facial emotion recognition using ResNet-50 pre-trained model for features extraction, Long Short Term memory network for classification using DISFA Plus data-set, and mapping of emotions into various learning affects. An accuracy of 95% is achieved on 6,081 samples of validation data set compared with the state- of- the art results. It is expected that this will cause an improvement to the online platform.

KEYWORDS

Online learning, Deep learning, facial emotion recognition, Learning Affect.

Functional Automation Testing of Palace Property Management Software

Divya Podduturi and Shahid Ali, Department of Information Technology, AGI Institute, Auckland, New Zealand

ABSTRACT

Till today, there are many web applications developed and tested so that end- user gets maximum user-friendly experience and satisfaction. For testing these applications and their functionalities manual testing is not always feasible. To overcome the challenge. Automation testing is used in which the testing process is automated with minimal manual intervention. Regression testing is deployed to perform the testing for routine flow of an application and when new features are added then automation testing becomes very useful. The focus of this study on automation testing. With the help of Selenium Webdriver scripts are executed and reports are generated. The default reports do not show any graphical representation of results. Hence, the main idea behind this study is to focus on automation testing with the combination of Selenium Webdriver, Maven, TestNG and Page Object Model which provides cutomised results to our study.

KEYWORDS

Automation testing, Regression testing, Visual Studio, C#, Selenium Webdriver, Agile- Scrum.

Privacy Preservation in Data Analytics using a Novel Algorithm Approach

Ms. Riya Gupta

ABSTRACT

To address the cutting edge protection dangers in information research by planning an adequate security saving information investigation procedure. The strategy applied is a non anonymized technique that utilizations combining semi identifiers and using differential security. The proposed method was applied to three informational indexes viz. Grown-up informative collection, Starlog informational index, and Indian Liver Patient informational index. All the informational indexes are uninhibitedly accessible in the UCI repository. The investigation presents "Blend Quasi Identifiers and applies Differential Privacy" (SQIDP), which ends up being a more proficient and versatile calculation. Contrasted with obscurity based algorithms, SQIDP is not inclined to comparability assaults, foundation information assaults, trait exposure, and derivation assaults. Anonymization, cryptographic, SWARM, and randomization strategies will lessen information utility, though SQIDP offers 100% information utility. Consequently, it is more proficient than different strategies. SQIDP is being applied on three distinctive informational indexes with 180, 853, and 24842 records, yet the algorithm execution time continued as before for each of the three informational indexes. SQIDP is a superior security protection strategy with 100% information utility since it is not anonymized. It keeps the proposal in numerous security enactments like the GDPR of Indias European Union PDP.

Quality Education as an Essential Tool for the Attainments of Sustainable Development Goals

Akinyooye Funmilola Elizabeth Ph. D and Aransi, Waliyi Olayemi, Department of Adult Education, Faculty of Education, University of Ibadan, Ibadan, Nigeria

ABSTRACT

There is no gainsaying in regarding provision of quality education as a major indication of a community’s social well-being, standard of living, social justice and above weapon for attaining national development. This is premised on the ground that education has come to playa very important role in the life of a man in particular and society in general, as it helps individual as a member of the society in the search for better life, better values, expansion of human domain of knowledge and serve as an instrument of recognition. It is inferred that education could be adjudged a powerful tool of enhancing intellectual abilities, shaping cultural attitudes and acquiring knowledge and skills.In the light of this, the paper examined the nexus between quality education and Sustainable Development Goals attainment. Literatures were scanned towards this direction and revealed that, ceteris paribus, the attainment of the SDGs which were prioritised in hierarchical order of ending poverty; eradicating hunger; ensuring good health and well-being of the people; ensuring quality education for all; achieving gender equality; ensuring provision of clear water and sanitation; ensuring affordable modern energy for all; promoting decent work and economic growth; creating industry innovation and infrastructure; reducing inequalities both within and among countries; achieving sustainable cities and communities; ensuring sustainable consumption and production pattern; controlling adverse climate action; sustaining life below water; protecting life on land, promoting peace, justice and strong institutions and lastly strengthening global partnership is a function of quality education. This indicated that provision and peoples’ access to quality education is both necessary and sufficient conditions for the attainment of the stated goals. It is therefore recommended that instruments vis-à-vis physical, human and financial resources needed to ensure availability and accessibility of quality education should be made available by the stakeholders of education. Also, both formative and summative evaluation of the school curricular and programmes should be done on regular basis. This would enhance to ascertain the extent to which the content deliver at various educational institutions work in harmony with the attainment of SDGs among others.

KEYWORDS

Quality Education, Attainments, Sustainable Development, Goals, Targets.

Classroom Acts on Low Literacy Adults Education Settings

Carlos Luís^1,2, Helena Afonso³ and Maria José Marcelino¹, ¹CISUC and FCTUC Universidade de Coimbra, Portugal, ²IEFP-Centro de Formação Profissional de Coimbra, Portugal, ³IEFP-Delegação Regional do Centro IEFP Coimbra; Portugal

ABSTRACT

This paper starts by discussing the relevance of dialogues in Adult Education and Training courses with low levels of literacy. In this group, the educational challenges are complex, and innovating the knowledge creation process involves a better understanding of the teaching/learning process. For such, three research questions have arisen: a) What types of Communicative Acts are linked to the teaching/learning process of adults with low literacy? b) Do all Communicative Acts contribute to this process? c) Are there any other types of acts that are present in the learning process? Based on a mixed methods, applied to a convenience sample, we used an ethnographic approach, and the Grounded Theory Methodology. The results showed that it was important to integrate the learners emotions in an existing framework, the SEDA Framework. We found also essential to expand the Communicative Acts coding, with a new set of 17 codes organized in 3 categories.

KEYWORDS

Adult Education, Literacy, Classroom acts, learning environments, Human-Computer-Interaction.

Engaging Students with Intellectual Disability in Stem Learning

Winnie Wing-mui So¹ and Qianwen He², ¹Department of Science and Environmental Studies, Education University of Hong Kong, ²Department of Science and Environmental Studies, Education University of Hong Kong

ABSTRACT

Recent studies show many of the current and future careers are having some form of STEM integration. And the new development of school STEM education provided opportunities for students to acquire the necessary knowledge, skills and attitude to face the challenges ahead. Involving students with different disabilities with STEM learning is of particular importance and usefulness in order not to deprive of their learning opportunities for the future. But how to expose students with intellectual disabilities (ID) to STEM learning has not been well explored. STEM learning can mean something different to each teacher and how they integrate it into their classroom. There are suggestions of using inquiry, engineering and technology to their advantages to support students with disabilities. In the design of STEM learning for students who have intellectual disabilities, teachers from a special school constructed a 4E model, which emphasizes inquiry and at the same time leverages technology and engineering to integrate learning content in a purposeful and informed way to better student engagement in lessons. Lesson observations were conducted to study students’ cognitive, affective and behavioural engagement in lessons. The lesson plans from teachers were analyzed to supplement the observation data to better understand the effect of lesson design with a focus on student engagement. Results showed that ID Students with mild ID in the classes responded and worked actively while students of moderate ID asked for more assistance. However, it is found that students were more engaged in engaging, exploring, and engineering, but less to be involved in explaining. This research provides the practical model and evidence to engage student engagement in STEM learning with a focus on inquiry, engineering and technology. This gives more insights for strategies in designing STEM learning for students with ID for better student engagement.

KEYWORDS

Students with Intellectual Disability, STEM learning, Student engagement, Inquiry, Technology, Engineering.

Social Media as Web 2.0 Technology Tool for Learning at the School Level in India

Ms. Bharti Kumari¹ and Dr. Parmod Kumar², ¹School of Education, Central University of Haryana, Haryana, India, ²School of Education, Central University of Haryana, Haryana, India

ABSTRACT

In the modern era of development, there are lots of challenges faced by developing countries. One of them is the technology used by learners, inappropriately and lack of knowledge. The main objective of the paper is to study the use of social media in school education, to find the applications of social media in school education, suggestions for future researches studies to be conducted. The present study is based on some research studies and literature reviews. The study found that social media is an effective tool for communication, collaboration, and interactive learning, also helps to improve the teaching and learning process, improving the imagination, the creative and innovative power of learners, and helpful in creating a technology-based school learning environment. The researcher also found that media literacy is very important for learners.

KEYWORDS

Social Media, Web 2.0 technology, School education.

International Higher Education Students Psychological Experiences of Covid-19: A Qualitative Study

Kayi Ntinda¹ and Nomazulu Ngozwana², ¹Department of Education Foundations and Management, Kwaluseni, Eswatini, ²Department of Adult Education, Kwaluseni, Eswatini

ABSTRACT

The Coronavirus pandemic 2019 (COVID -19) is hastily spreading, bringing pressure and challenges to international students in higher education institutions who were locked down on campus during the COVID-19 outbreak. We sought to explore psychological experiences of international higher education students during the COVID-19 pandemic in Eswatini. A qualitative phenomenological approach was adopted. Participants were 15 conveniently selected international higher education students who were locked down on Campus from March 20 to September 25, 2020. The interviews were conducted face to face. Data were thematically analyzed. The study was informed by the ecological systems theory. The psychological experiences of international higher education students during COVID-19 pandemic were categorized in to four themes. Frist negative emotions present in early stages involving discomfort, anxiety and helplessness caused by isolation, fear and concern for safety. Second self-coping strategies included psychological and life adjustments, social distancing, acceptance of wearing of masks, hand sanitizing and online learning. Third we reported growth under the crisis which included affection for family members, peers and self-reflection. Finally, we reported that positive emotions occurred concurrently with negative emotions. During the COVID-19 crisis positive and negative emotions of international higher education students intertwined and co-occurred. Self-coping strategies and psychological growth played a crucial role in maintaining mental health of the students. The international students in higher education were resilient in coping with the COVID-19 crisis and lockdown situation which enhanced their participation in online learning.

KEYWORDS

International higher education students, psychological experiences, COVID-19, Eswatini.

Compulsory Education in Uzbekistan

Jakhongir Shaturaev, Assistant Professor of the Department of “Corporate Economics and Business Analytics”, Tashkent State University of Economics, Islam Karimov avenue, 49 100003 Tashkent, Uzbekistan

ABSTRACT

Locating in the heart of Central Asia, Uzbekistan pays out an enormous portion of its budget and attention to compulsory education in the area. Meanwhile, public education is afflicted by several issues apart from excellence both in the teaching and learning process. The author tried to define the current circumstance of primary education and sought possible solutions for them. Through field surveys and data analysis methods used throughout the investigation. Found data shows that improper infrastructure of government expenditure on education, low salary, and limited quota in pedagogical universities lead to a shortage of teachers in rural areas. The government of Uzbekistan needs further educational reforms in the area of public education, teacher training, and re-training programs, and increasing teachers’ salaries.

KEYWORDS

compulsory education, primary schools, reforms, teaching-learning process, Uzbekistan.

The Evaluation Case Study of Online Course During Pandemic Period in Mongolia

Uranchimeg Tudevdagva^1,2 and Bazarragchaa Sodnom³ and Selenge Erdenechimeg³, ¹Faculty of Computer Science, Technical University of Chemnitz, Germany ²Power Engineering School, Mongolian University of Science and Technology, Mongolia, ³Mongolian University of Pharmaceutical Sciences, Mongolia

ABSTRACT

This paper describes test and case study of self evaluation of online courses during pandemic time. Due to the Covid-19 the whole world needs to sit on lockdown in different periods. Many things need to be done in all kind of business including education sector of countries. To sustain the education development teaching methods had to switch from traditional face to face teaching to online courses. The government made decisions in short time and educational institutions had no time to prepare the materials for the online teaching. All courses of Mongolian University of Pharmaceutical Sciences switched to online lessons. Challenges were raised before professors and tutors during online teaching. Our university did not have a specific learning management system for online teaching and e-learning. Therefore professors used different platforms for their online teaching such as Zoom, Microsoft teams for instance. Moreover, different social networking platforms played active role for communication between students and professors. The situation is very difficult for professors and students. To measure the quality of online courses and to figure out positive and weak points of online teaching we need evaluation of e-learning. The focus of this paper is to share the evaluation process of e-learning based on structure oriented evaluation model.

KEYWORDS

Key target, sub target, SURE model, evaluation of evaluation, Mongolian university of pharmaceutical sciences.

A New Normal: Artificial Intelligence in Education

Deepika Gahlawat, Central University of Haryana, Haryana, India

ABSTRACT

A significant part of the modern era is our tenacious adoption of technology. A swing in demography and technical shift is on the road to the use of Artificial intelligence.major displacement andeconomic transition to digital globalization make it necessary for the young generation to succeed in the field. Teaching artificial intelligence adds to the imagination, critical reasoning, expertise in innovation, problem-solving and collaborative skills. It also contributes to the building of life and career skills. The core objective of the paper is to explore what is Artificial intelligence and the reason to introduce the same at the school level at the earliest possible.

KEYWORDS

Artificial Intelligence (AI), Critical Thinking, problem-solving.

Mapping out O*NET data to inform workforce readiness certification programs

Micheline Al Harrack, Workforce Development: Cybersecurity Education, Training, and Certification

ABSTRACT

The Occupational Information Network O*NET did not explore possible uses of O*NET data to inform workforce readiness certification programs. The O*NET database will be used to map out education requirements and how they relate to professional certifications as required by the employers and job designers in accordance with the National Initiative for Cybersecurity Careers and Studies NICCS. The search will focus on the “Information Security Analysts” occupation as listed on O*NET, careeronestop, U.S. Bureau of Labor Statistics, and finally tied back to NICCS source work role to identify certifications requirements.

A New Lidar-Based Approach for Poles and Distribution Lines Detection and Modelling

Mohamed Gaha¹, Wael Jaafar², Jed Fakhfekh³, Guillaume Houle¹, Jihene Ben Abderrazak³ and Michel Bourgeois¹, ¹Hydro-Quebec Research Institute, Quebec, CANADA, ²Carleton University, Ontario, CANADA, ³ESPRIT Engineering School, Tunis, TUNISIA

ABSTRACT

Vegetation is the major cause of overhead power line failures. According to a recent Hydro-Quebec analysis, more than 60% of the power outages are related to vegetation. Specifically, when branches/trees nearby the distribution network interact with extreme weather conditions, e.g., melting snow and heavy rain, they may bend and cause power outages. To ensure the reliability of our distribution network, millions of dollars are yearly spent for pruning trees and trimming branches. Aiming to reduce these costs, we recently adopted a new approach based on light detection and ranging (LiDAR) data. Indeed, we scanned 150 km of Hydro-Quebec’s network using a mobile LiDAR system. Through data analysis, we target automatic detection of hot spots, i.e., locations of threatening branches to distribution lines. However, such an operation cannot be accurately completed without a prior efficient detection of poles and lines locations, even for incomplete or missing LiDAR data. Hence, we propose here a low-complex and robust method for poles/distribution lines detection and lines modelling. Through customized filtering and detection, we efficiently detect poles and distribution lines with high accuracy and recall. Indeed, poles are detected with an accuracy of 94.5% and a recall of 89.7%, while lines are detected with an accuracy of 84% and a recall of 98.9%. Finally, our approach reconstructs power lines with a distance deviation from the real ones below 20 cm, in 89% of the cases. Such accuracy is required to automatically evaluate the closeness of vegetation to distribution lines and prevent power outages.

KEYWORDS

Mobile LiDAR, power lines, distribution lines detection, poles detection, distribution lines modelling.

Can Existing Image Quality Metrics Assess Enhanced Images?

Gehang Huang¹ and Xinwei Liu², ¹HDU-ITMO Joint Institute, Hangzhou Dianzi University, Hangzhou, China, ²College of Big Data and Software Engineering, Zhejiang Wanli University, Ningbo, China

ABSTRACT

Most of image quality metrics are designed for degraded images, and their performance are highly correlated with the human visual system. The rapid development of image enhancement techniques raises a new issue: how to evaluate the performance of image enhancement methods? It is obvious that many benefits can be obtained if existing image quality metrics can be applied to the performance evaluation of image enhancement algorithms. However, quite a few studies have investigated such an issue. In this paper, we conduct psychophysical experiments to investigate the performance of existing image quality metrics on enhanced images. Two databases that specifically designed for enhanced images are selected in the experiments for the collection of subjective evaluation results. The objective evaluation results are obtained from six selected image quality metrics. By calculating the correlation between subjective and objective results we can answer the question: the performance of existing selected image quality metrics is low on enhanced images.

KEYWORDS

Image quality metric, image enhancement, psychophysical experiment, performance evaluation.

Modelling Demand for Outpatient Health-care Services in Kenya using Artificial Neural Networks

Assumpta Mbatha, Kennedy Ogada and Tobias Mwalili, Department of Computing and Information Technology, Jomo Kenyatta University of Agriculture and Technology, Nairobi, Kenya

ABSTRACT

Artificial neural networks (ANNs) have been utilized extensively in various engineering and science fields as they are models that are capable of incorporating both nonlinear and linear effects in the modeling process. This study extends this modeling technique to modeling outpatient health care demand to help understand, determine and predict the nature of outpatient health care provision in Kenya. The cost of outpatient health-care services and the number of patients sourcing this services were evaluated as the determinants of demand in which the cost of outpatient health-care proved to be a significant factor in the estimation and prediction of outpatient health-care services in Kenya. The logistic regression and the artificial neural network architecture were used in the modeling of demand for outpatient health-care services in which training of the data before model fitting helped reduce the bias associated with estimating model parameters.

KEYWORDS

Artificial Neural Networks, Outpatient, Logistic Regression, Health-care services.

Fabric Defect Detection Based on Faster RCNN With CBAM

Yuan He, Han-Dong Zhang, Xin-Yue Huang, Francis Eng Hock Tay, Department of Mechanical Engineering, National University of Singapore, Singapore

ABSTRACT

In the production process of fabric, defect detection plays an important role in the control of product quality. Consider that traditional manual fabric defect detection method are time-consuming and inaccuracy, utilizing computer vision technology to automatically detect fabric defects can better fulfill the manufacture requirement. In this project, we improved faster rcnn with convolutional block attention module (cbam) to detect fabric defects. Attention module is introduced from graph neural network, it can infer the attention map from the intermediate feature map and multiply the attention map to adaptively refine the feature. This method improve the performance of classification and detection without increase the computation-consuming. The experiment results show that faster rcnn with attention module can efficient improve the classification accuracy.

KEYWORDS

Fabric defects detection, Faster RCNN, Convolutional block attention module, Deep learning.

Fabric Defect Detection Based on Object as Point

Yuan He, Xin-Yue Huang, Francis Eng Hock Tay, Department of Mechanical Engineering, National University of Singapore, Singapore

ABSTRACT

In the field of fabric manufacturing, many factories still utilise the traditional manual detection method. It requires a lot of labour, resulting in high error rates and low efficiency. In this paper, we represent a real-time automated detection method based on object as point. This work makes three attributions. First, we build a fabric defects database and augment the data to training the intelligence model. Second, we provide a real-time fabric defects detection algorithm, which have potential to be applied in manufacturing. Third, we figure out centernet with soft nms will improved the performance in fabric defect detection area, which is considered an nms-free algorithm. Experiment results indicated that our lightweight network based method can effectively and efficiently detect five different fabric defects.

KEYWORDS

Fabric defects detection, Object as point, data augmentation, Deep learning.

Analysis of parameters used for measuring performance of mobile applications

K. Durga Sowjanya, P. Bhaskara Reddy, Department of Electronics and Communication Engineering, MLR Institute of Technology, Hyderabad, India

ABSTRACT

To make a mobile application more reliable, performance is the most important parameter. Resource management plays a crucial role inside the development of mobile application. The better the app performs, the better the chances are for the app in the market. Careful attention should be required for number of non-functional requirements to achieve ultimate goal of mobile application development. Mobile application and its impact on market are analyzed by performance evaluation process. For getting better performance, efficient utilization of physical resources is required. These resources provide actionable information to the application developer for optimizing performance and increases efficiency of mobile application during development cycle. Due to limited processing power and memory resources, smart phones and tablets keep the hardware components always in running states. But slow applications drain the batteries of their devices. To maximize the utilization of resources, developers should consider hardware limitations too which make an effort to optimize performance and efficiency of the application. This paper presents a qualitative performance analysis of mobile applications. This paper describes the factors considered for analyzing performance and corrective measures to achieve better performance.

KEYWORDS

Mobile application, Resource management, Performance evaluation, optimizing performance.