Supporting University Students During the Pandemic: A Study on The Efficacy of a Mentalizing Online Group Counselling

Background: University counselling services assume a fundamental support function for students who are facing moments of crisis during their academic career. Such services often aim to reduce drop-out rates and achieve improvement in terms of psychological well-being. COVID-19 contagion containment measures have also had an impact on the psychological health of university students and their ability to cope with important developmental tasks. It has become necessary, therefore, to offer online counselling services which has become, however, the means of choice to support students during the university course in the pandemic era, as a complementary intervention to the traditional face-to-face approach. Methods: In a clinical and health psychology perspective, this study aims to analyze the efficacy of 13 online counselling groups involving 66 underachieving students, lagging with their studies. The intervention has adopted the methodology of the Narrative Mediation Path, which aims at promoting mentalization, academic engagement and psychological well-being in order to have an impact on students' academic performance and prevent university dropouts. At the beginning and end of counselling the following measures were administered: a) Reflective Functioning Questionnaire, b) Psychological General Well-Being Index Short Form, c) Academic Performance Inventory, d) University Student Engagement Inventory, e) Group Climate Questionnaire. Results: The results showed that online counselling groups enabled an overall improvement in all the variables considered. Conclusion: Overall, the present study showed the efficacy of the online group counselling service in supporting students during the pandemic period and in coping with the difficulties encountered during the academic career © 2023 by the Author(s);licensee Mediterranean Journal of Clinical Psychology, Messina, Italy. This article is an open access article, licensed under a Creative Commons Attribution 4.0 Unported License

Removing inclusion: An analysis of exclusionary processes in the Italian school system

Since the abolition of special schools in 1971, Italy has been depicted as enjoying a fully inclusive education system, where all children are welcome to attend mainstream schools in spite of disability or other 'special' conditions. However, research shows that inequality, segregation and exclusion are still deeply rooted in the Italian school environment, usually taking the form of micro-exclusion strategies that prevent the full access and participation of students from a minority or underprivileged background. Adopting an intersectional approach, the chapter examines how exclusionary policies and practices currently affect compulsory education in Italy by analysing three relevant examples: the dynamics of early school leaving, the evolution of the special education needs construct, and the marginalisation of disadvantaged students in the recent Covid-19 pandemic. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023. All rights reserved.

Management of the Surgical Patient during the First and Second Wave of COVID-19 Pandemic: Experience of a Tertiary Care Centre

Objective: Due to the COVID 19 pandemic healthcare providers all over the world had brought some changes in the management of surgical patients. This study is aimed to estimate the impact of pandemic on surgical practices. Material(s) and Method(s): We conducted a retrospective review of the medical records of all patients admitted to the department of general surgery (both elective & emergency), SCB Medical College and Hospital, Odisha, India from April 1 to July 31, 2020, and 2021 and the records were those of patients who were admitted in the same period in 2019. Data collection includes the number of admissions, the reason for admission, the age & gender of the patients admitted patients and type of management. Result(s): There was a 57.5% reduction in total admission during first COVID in pandemic 2020 and 58.7% reduction during second wave of pandemic in 2021. The proportion of patient presenting to emergency department was more in 2020 and 2021 than 2019. Number of emergency admission decreased by 46.54% in 2020 and 46% in 2021. There was a 79.5% drop in the number of out-patients admission in 2020 and 84% in 2021. Furthermore a 79.8% reduction in elective surgical intervention noticed in 2020 and 80% in 2021. Conservative management was preferred over surgical management during the COVID era. Conclusion(s): COVID-19 has led to a drastic reduction in outpatient and elective surgical practices. Hence creating a major concern for all surgeons about the critical situation.Copyright © 2023, Dr Yashwant Research Labs Pvt Ltd. All rights reserved.

Prediction of Dropout Students in Massive Open Online Courses Using Ensemble Learning: A Pilot Study in Post-COVID Academic Session

High dropout rate is a critical problem in MOOCs. The prime objective of this study is to identify possible dropout students at the early stage of the course and reducing the number of dropouts providing proper feedback to address the relevant factor. A prediction model based on stacking ensemble machine learning is proposed to identify whether a learner is at risk of dropping a course. The proposed stacked ensemble model outperformed with an accuracy of 93.4% compared to other popular machine learning classifiers. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Academic Analytics Applied in the Study of the Relationship Between the Initial Profile of Undergraduate Students and Early Drop-Out Rates. Defining the Variables of a Predictor Instrument

The field of university dropout research is of utmost importance especially in the current context arising from the Covid-19 pandemic. Students who started their degrees in the last two years completed their pre-university studies during various phases of confinement and by combining traditional and virtual training. In this scenario, students' motivation and the way they cope with the difficulties of their first year of university are very relevant and will depend on a multitude of personal and social variables in their immediate environment. Previous studies have shown that many university students drop out of their studies early, but what factors and to what extent they affect this dropout is still a field under study. This paper focuses on the identification, classification and evaluation of a set of indicators based on teacher and tutor perception in different fields of study by applying quantitative and qualitative techniques. The results of pilot studies developed support the approach adopted, as they show how teachers can identify students at risk of dropping out at the beginning of the course and take proactive measures to monitor and motivate them, thus reducing the possibility of dropout. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

COVID-19 Detection from Pulmonary CT Images using Neural Networks based on Dropout-Driven Hidden Layers

Enhanced diagnosis with considerably good sensitivity and specificity is highly indispensable for COVID-19 diagnosis using radiological data to combat hazardous viral infection. Accuracy of diagnosis is a very important part that helps in further triaging and disease management. Artificial intelligent techniques using Convolutional Neural Networks and their modified alternatives have been recognized to be the salvation in chaotic situations and emergencies. Despite their immense ability to give quality results, they suffer from overfitting problems which have to be reduced by regularizing the networks. Dropout is one such regularization that modifies the network to achieve improved performance by discarding the unwanted nodes in the network layers. A simple neural network architecture inspired by former renowned architectures with dropout-driven hidden layers, CVDNN is built and experimented with for various dropout probabilities (0.1, 0.25, 0.5 and 0.75). The model was also tested with different numbers of dense layers: CVDNN1 with a single dense layer and CVDNN2 with two dense layers of a fixed dropout probability of 0.5 in it. The models are trained and tested with pulmonary computed tomography images to distinguish COVID-19 abnormality against normal cases. The CVDNN2 model presents better functioning with improved performance measures than CVDNN1 with an accuracy of 92.86 % accuracy, 90.21% sensitivity and a specificity of 95.52% for the dataset used. Dropout probabilities of 0.25 and 0.5 present reliable and better results compared to the other values experimented with. Hence a dropout-driven hidden layer can enhance the neural network's performance by choosing either 0.25 or 0.5 preferably for different applications. © 2023 IEEE.

Ruptured school trajectories: understanding the impact of COVID-19 on school dropout, socio-emotional and academic learning using a longitudinal design.

This paper presents the findings of longitudinal research conducted in Ethiopia exploring the effects of COVID-19 school closures on children's holistic learning, including their socio-emotional and academic learning. It draws on data from over 2,000 pupils captured in 2019 and 2021 to compare primary school children's dropout and learning before and after school closures. The study adapts self-reporting scales used in similar contexts to measure grade 4-6 pupils' social skills and numeracy. Findings highlight the risk of widening inequality regarding educational access and outcomes, related to pupils' gender, age, wealth and location. They also highlight a decline in social skills following school closures and identify a positive and significant relationship between pupils' social skills and numeracy over time. In conclusion, we recommend a need for education systems to promote children's holistic learning, which is even more vital in the aftermath of the pandemic.

Engagement over telehealth: comparing attendance between dialectical behaviour therapy delivered face-to-face and via telehealth for programs in Australia and New Zealand during the Covid-19 pandemic.

BACKGROUND: While the COVID-19 crisis has had numerous global negative impacts, it has also presented an imperative for mental health care systems to make digital mental health interventions a part of routine care. Accordingly, through necessity, many Dialectical Behaviour Therapy (DBT) programs transitioned to telehealth, despite little information on clinical outcomes compared with face-to-face treatment delivery. This study examined differences in client engagement (i.e. attendance) of DBT: delivered face-to-face prior to the first COVID-19 lockdown in Australia and New Zealand; delivered via telehealth during the lockdown; and delivered post-lockdown. Our primary outcomes were to compare: [1] client attendance rates of DBT individual therapy delivered face-to-face with delivery via telehealth, and [2] client attendance rates of DBT skills training delivered face-to-face compared with delivery via telehealth. METHODS: DBT programs across Australia and New Zealand provided de-identified data for a total of 143 individuals who received DBT treatment provided via telehealth or face-to-face over a six-month period in 2020. Data included attendance rates of DBT individual therapy sessions; attendance rates of DBT skills training sessions as well as drop-out rates and First Nations status of clients. RESULTS: A mixed effects logistic regression model revealed no significant differences between attendance rates for clients attending face-to-face sessions or telehealth sessions for either group therapy or individual therapy. This result was found for clients who identified as First Nations persons and those who didn't identify as First Nations persons. CONCLUSIONS: Clients were as likely to attend their DBT sessions over telehealth as they were face-to-face during the first year of the Covid-19 pandemic. These findings provide preliminary evidence that providing DBT over telehealth may be a viable option to increase access for clients, particularly in areas where face-to-face treatment is not available. Further, based on the data collected in this study, we can be less concerned that offering telehealth treatment will compromise attendance rates compared to face-to-face treatment. Further research is needed comparing clinical outcomes between treatments delivered face-to-face compared delivery via telehealth.

Dupchanchia Model of Students' drop-out control through engagement and appreciation after Covid-19: A Behavioral Policy intervention in the field administration of Bangladesh

Bangladesh is a country which has successfully accomplished the millennium development goals. Upon such accomplishment and with new growing consensus with the global community, the country at present is in pursuit of achieving the sustainable development goals with mostly concentrating on the education sector. However, what impacted the growth and pace of the initiatives was the overwhelming impact of Covid-19 and the lockdown afterwards. Academic institutions remained closed at least for two years which resulted in a compromising number of students after the resumption. This study starts from the identification of a genuine problem with original field level data of the gradually declining number of students. It requires policy intervention centrally and locally. Government of Bangladesh has been deploying some traditional method like vocational and stipend system which is involved with large amount of monetary disbursement. This study found that government like Bangladesh should introduce new or customized education policy which can reduce budgetary involvement and change the choice structure of students. This approach and method has the transferability whereby other similar states can adapt. As a crucial part of local government, I have been working as the chief executive officer of a sub-district called Dupchanchia, and coordinating government departments to implement government policy. After rigorous discussion and brainstorming among the local stakeholders and teachers we uncovered that the students have become demotivated, traumatized and panicked of social engagement and any form of shared activities like classes, games and others. It required us to find out a local policy solution followed by a detailed literature review and primary data collection maneuver. Taking twenty schools into consideration for the study, the project initiated a behavioral policy intervention in ten particular schools and did not interfere with the other ten schools. I engaged local teachers, students and other related stakeholders and continued to use six behavioral tools to change the choice structure of the students of ten selected schools. We observed other ten schools without intervening in their environment and academic atmosphere at all. At the end of the study we collected data through key informant interviews and focus group discussion engaging teachers, peoples' representatives and government officials. Behavioral public policy intervention like nudge and engagement approaches are found to have a positive relation with the change in students number and their performance in the academic and co-curricular activities. This approach may contribute to controlling students' drop out in the lower and lower middle income countries after Covid shock. This policy intervention may have some challenges and limitations which need intensive and rigorous pre-study and prolonged design. Nevertheless, it has unlimited opportunities to be addressed. [ FROM AUTHOR] Copyright of Technium Social Sciences Journal is the property of Technium Press Constanta and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full . (Copyright applies to all s.)

Forecasting Graduation Schedule Model of Higher Education Learners using Feature Selection Techniques

The biggest concern for learners in the 21st century is that graduation does not fit their educational plan. This research aims to study factors affecting students' academic achievement in the Faculty of Education during the COVID-19 crisis in Thailand. The data was on the academic achievement of 90 students from the Bachelor of Art Program in Educational Technology and Communications at the Faculty of Education, Naresuan University, Phitsanulok, Thailand. The research process was analyzed using data mining techniques, including CRISP-DM procedures, decision tree algorithm, forward selection analysis, cross-validation techniques, and confusion matrix performance. This research found that the course 001211 Fundamental English was the most significant subject for delayed graduation, where the developed model has a very high level of accuracy (89.00%). Researchers can use such a model to create effective planning strategies for preventing graduation failures. © 2023 Seventh Sense Research Group®

School dropout of university students in the career of Law and Social Sciences

The COVID-19 pandemic has been a great challenge for education at all levels, teachers and students have had to adapt to the new online class modality, despite the economic, technological and pedagogical obstacles that have arisen. The objective of this research is to identify the reasons that students of the Law and Social Sciences career at a public university in northwestern Mexico had to abandon their studies temporarily or permanently during the 2020-2021 school year, in times of pandemic by COVID-19. A study was carried out from the qualitative approach, of a descriptive type, the technique used was the semi-structured and indepth interview. Of a total of 32 students who had dropped out of school, 19 agreed to participate in the study. Although it is true that the virtual modality or online classes represented a lifeline for students in times of COVID, those who could not or did not want to adapt to this new school environment left their studies, the pedagogical aspect being the main reason why they were unable to adapt to the online work modality since they did not have adequate conditions at home © 2023, Revista de Ciencias Sociales.All Rights Reserved.

TFOS Lifestyle: Impact of contact lenses on the ocular surface.

Several lifestyle choices made by contact lens wearers can have adverse consequences on ocular health. These include being non-adherent to contact lens care, sleeping in lenses, ill-advised purchasing options, not seeing an eyecare professional for regular aftercare visits, wearing lenses when feeling unwell, wearing lenses too soon after various forms of ophthalmic surgery, and wearing lenses when engaged in risky behaviors (e.g., when using tobacco, alcohol or recreational drugs). Those with a pre-existing compromised ocular surface may find that contact lens wear exacerbates ocular disease morbidity. Conversely, contact lenses may have various therapeutic benefits. The coronavirus disease-2019 (COVID-19) pandemic impinged upon the lifestyle of contact lens wearers, introducing challenges such as mask-associated dry eye, contact lens discomfort with increased use of digital devices, inadvertent exposure to hand sanitizers, and reduced use of lenses. Wearing contact lenses in challenging environments, such as in the presence of dust and noxious chemicals, or where there is the possibility of ocular trauma (e.g., sport or working with tools) can be problematic, although in some instances lenses can be protective. Contact lenses can be worn for sport, theatre, at high altitude, driving at night, in the military and in space, and special considerations are required when prescribing in such situations to ensure successful outcomes. A systematic review and meta-analysis, incorporated within the review, identified that the influence of lifestyle factors on soft contact lens dropout remains poorly understood, and is an area in need of further research. Overall, this report investigated lifestyle-related choices made by clinicians and contact lens wearers and discovered that when appropriate lifestyle choices are made, contact lens wear can enhance the quality of life of wearers.

The use of Deep Learning techniques in E-Learning systems and MOOCs

E-Learning and Massive Open Online Courses are old techniques, but since the Coronavirus, they have become more popular again. Students already suffer from a lack of concentration and motivation in traditional courses;thus, this lack affects online courses. Furthermore, another important Online Learning systems problem is the difference between learners in terms of Learning Styles, abilities, social characteristics as well as preferences, background, and other psychological and mental features. Generally, these features are not taken into account by scientists. Therefore, Deep Learning techniques and Datasets have been used to improve E-Learning systems and MOOCs in several aspects such as: predicting dropout, Learning Styles and performance of online learners, and even their attention after taking an online course. In this work, we have studied and analyzed many recent works in the area of using Deep Learning techniques to improve Online Learning systems and MOOCs. This analysis shows what researchers rely on to improve E-Learning and MOOCs and demonstrates that research does not use the definition of the appropriate Learning Style frequently. However, the most used ones are dropout and performance of learners. In another hand, learners' attention is still gap. © 2022 IEEE.

Learning Analytics Overview: Academic Approach and Machine Learning Possibilities

In the Covid-19 era, traditional lecture-based teaching has been undergoing changes in learning design, learners' engagement, and technology integration. Online learning has become an integral part of education around the globe due to its flexibility in learning with respect to place and time. These online courses are available to larger audiences and enable students to have more freedom over the study process. However, freedom also means that instructors have less control to keep students making progress on the course. The flexibility of online courses is encouraging the students to enrol with a few clicks but most of these students are dropping out due to losing interest in the course contents within a few weeks. On the contrary, online learning produces large amounts of data that can be used to follow the learning process and give useful insights for both teachers and students. Learning analytics algorithms utilize these data to identify the factors and parameters that can explain learners' dropout rate, learning performance, as well as suggest possible actions to intrigue learners' active engagement. To conduct further research on the parameters affecting learners' participation in an online course, it is essential to find out the previous research works, best practices, and research trends of learning analytics. In the scope of this work, the authors formulated a search query to generate a pool of most relevant papers from the Scopus database and, hence, identified seven clusters which illustrate the landscape of the learning analytics domain. The authors also employed the Latent Dirichlet allocation (LDA) topic modeling algorithm which is a form of text data mining and statistical machine learning approach to compare the similarities with the clusters generated as a part of literature review. The authors further analyzed the papers in each cluster to identify the parameters which are significant to build a predictive model on learners' dropout rate. This work not only provides a baseline to conduct further research to find out the parameters affecting learners' retention rate but also introduces a systematic methodology to validate the findings of the literature review with a data-driven algorithmic approach. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2022.

Design of ear-contactless stethoscope and improvement in the performance of deep learning based on CNN to classify the heart sound.

Cardiac-related disorders are rapidly growing throughout the world. Accurate classification of cardiovascular diseases is an important research topic in healthcare. During COVID-19, auscultating heart sounds was challenging as health workers and doctors wear protective clothing, and direct contact with patients can spread the outbreak. Thus, contactless auscultation of heart sound is necessary. In this paper, a low-cost ear contactless stethoscope is designed where auscultation is done with the help of a bluetooth-enabled micro speaker instead of an earpiece. The PCG recordings are further compared with other standard electronic stethoscopes like Littman 3 M. This work is made to improve the performance of deep learning-based classifiers like recurrent neural networks (RNN) and convolutional neural networks (CNN) for different valvular heart problems using tuning of hyperparameters like learning rate of optimizers, dropout rate, and hidden layer. Hyper-parameter tuning is used to optimize the performances of various deep learning models and their learning curves for real-time analysis. The acoustic, time, and frequency domain features are used in this research. The investigation is made on the heart sounds of normal and diseased patients available from the standard data repository to train the software models. The proposed CNN-based inception network model achieved an accuracy of 99.65 ± 0.06% on the test dataset with a sensitivity of 98.8 ± 0.05% and specificity of 98.2 ± 0.19%. The proposed hybrid CNN-RNN architecture attained 91.17 ± 0.03% accuracy on test data after hyperparameter optimization, whereas the LSTM-based RNN model achieved 82.32 ± 0.11% accuracy. Finally, the evaluated results were compared with machine learning algorithms, and the improved CNN-based Inception Net model is the most effective among others.

Community College Student Persistence During the COVID-19 Crisis of Spring 2020

Objective: This quantitative study examines the impact of the COVID-19 pandemic on students' persistence at a minority-serving, open-access, public, urban community college in New York City. Specifically, the project looked at factors associated with mid-semester college withdrawals during spring 2020 when the college shifted to remote instruction due to the COVID-19 pandemic. Method: Utilizing data from three spring semesters (spring 2018, 2019, and 2020), four logistic regression models tested the marginal effects of student background and college program factors on mid-semester withdrawal and the moderating effect of spring 2020, the COVID-19 outbreak semester. Results: Findings indicated that the withdrawal rates were higher for new students, men, minoritized students, and part-time students across all three spring semesters. Spring 2020 disproportionally affected part-time students, men, Black students, as well as readmitted students. The greatest increase in the probability of mid-semester college withdrawal was observed for Black men who had been enrolled part-time in spring 2020. Belonging to a highly structured full-time study program protected students from leaving mid-semester, although this protection was weaker in spring 2020 and spring 2019 compared to spring 2018. Contributions: The research highlights the equity gap for Black men at the college and points to additional factors contributing to mid-semester college attrition. The work provides insights into factors that worsened during the COVID-19 pandemic. The study thereby contributes to understanding short-term risk factors for vulnerable student populations and adds to the body of literature on crisis situations in higher education. © The Author(s) 2023.

A graph-based edge attention gate medical image segmentation method

For the purpose of solving the problems of missing edges and low segmentation accuracy in medical image segmentation, a medical image segmentation network (EAGC_UNet++) based on residual graph convolution UNet++ with edge attention gate (EAG) is proposed in the study. With UNet++ as the backbone network, the idea of graph theory is introduced into the model. First, the dropout residual graph convolution block (DropRes_GCN Block) and the traditional convolution structure in UNet++ are used as encoders. Second, EAGs are adopted so that the model pays more attention to image edge features during decoding. Finally, aiming at the imbalance problem of positive and negative samples in medical image segmentation, a new weighted loss function is introduced to enhance segmentation accuracy. In the experimental part, three datasets (LiTS2017, ISIC2018, COVID-19 CT scans) were used to evaluate the performances of various models;multiple groups of ablation experiments were designed to verify the effectiveness of each part of the model. The experimental results showed that EAGC_UNet++ had better segmentation performance than the other models under three quantitative evaluation indicators and better solved the problem of missing edges in medical image segmentation. © 2023 The Authors. IET Image Processing published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology.

Interrupted trajectories: university students' reasons for temporarily suspending their studies during the pandemic

The Covid-19 pandemic has affected the continuity of the educational trajectories of many students who have had to suspend their studies temporarily or permanently. In the case of the National Autonomous University of Mexico (unam), the General Secretary's office stated in 2020 that 20% of students were at risk of dropping out of school due to the pandemic, while by 2021 the number of students requesting temporary suspension of their studies increased by 228%. In order to identify the reasons put forward by the student body for submitting to the Technical Council of their entity a temporary suspension of studies, a mixed convergent study was conducted in which the causes given by 268 students of the Psychology degree of the Faculty of Psychology of the unam, both open and attendance systems, for temporarily suspending their studies are analyzed. The purpose is to understand, based on the voice of the students, the problems they have been experiencing during the pandemic without being able to achieve continuity with the theoretical trajectory typified by the school administration. The results indicate the interrelation of factors associated with socioeconomic or labor conditions of the students and their families, responsibilities and problems at home, lack of appropriate technological infrastructure, health and emotional problems, educational conditions and prior academic lag, as well as dissatisfaction and demotivation regarding the online education received and the reduction of the school semester. Proposals for attention and prevention of college dropout during the pandemic are discussed. © 2022 Universidad Nacional Autonoma de Mexico. All rights reserved.

Causes of academic dropout in higher education in Andalusia and proposals for its prevention at university: A systematic review

Introduction: One of the main problems facing the university system is the high student dropout rate due to a number of variables, accentuated by the COVID-19 pandemic. This is a problem not only in Spanish universities but is prevalent worldwide. It is therefore important to understand and analyze the underlying reasons for dropout so that it can be addressed and mechanisms implemented to limit dropout in higher education to the greatest extent possible. Method: A systematic review was carried out summarizing the results of studies and reports on university dropout in Spain and specifically in the universities of the Autonomous Community of Andalusia. The review was conducted in accordance with the PRISMA statement by searching the scientific databases Scopus and Web of Science, limiting the search to articles published between 2010 and 2022. Results: The main publications in both Spain and the Autonomous Community of Andalusia were identified. The review included the main causes of university dropout indicated in each of the selected studies and the proposals to reduce it, including educational policies, the rise of distance education, academic failure in basic educational stages, and social, personal, psychological, and economic variables. Conclusion: There is a lack of research on university dropout, with only 25% of Spanish universities having carried out research on this subject in the last 12 years. The studies analyzed conclude that the most frequent causes of university dropout are associated with low academic performance, poor social support in the new academic environment, low socio-economic status, pessimism, and lack of motivation, together with other less significant factors such as poor relationships with teachers, lack of vocation, work incompatibility, and previous academic performance. Further research on the causes of university dropout and its prevention is needed both before university entrance, by providing meaningful information to secondary school students, and during the university stay, through institutional and teaching policies that improve family support and social roots, produce positive academic experiences, favor associationism, and encourage activities that improve planning and time management, together with cognitive learning strategies, motivational strategies and the use of advanced learning materials [such as Information and Communication Technology (ICT) tools]. Copyright © 2023 de la Cruz-Campos, Victoria-Maldonado, Martínez-Domingo and Campos-Soto.

Improving Radiology Report Generation via Object Dropout Strategy and MLP-based Captioner

Recently, automatically generating radiology reports has been addressed since it can not only relieve the pressure on doctors but also avoid misdiagnosis. Radiology report generation is a fundamental and critical step of auxiliary diagnosis. Due to the COVID-19 pandemic, a more accurate and robust structure for radiology report generation is urgently needed. Although radiology report generation is achieving remarkable progress, existing methods still face two main shortcomings. On the one hand, the strong noise in medical images usually interferes with the diagnosis process. On the other hand, these methods usually require complex structure while ignoring that efficiency is also an important metric for this task. To solve the two aforementioned problems, we introduce a novel method for medical report generation, the termed attention-guided object dropout MLP(ODM) model. In brief, ODM first incorporates a tailored pre-trained model to pre-align medical regions and corresponding language reports to capture text-related image features. Then, a fine-grained dropout strategy based on the attention matrix is proposed to relieve training pressure by dropping content-irrelevant information. Finally, inspired by the lightweight structure of Multilayer Perceptron(MLP), ODM adopts an MLP-based structure as an encoder to simplify the entire framework. Extensive experiments demonstrate the effectiveness of our ODM. More remarkably, ODM achieves state-of-the-art performance on IU X-Ray, MIMIC-CXR, and ROCO datasets, with the CIDEr-D score being increased from 26.8% to 41.4%, 21.1% to 30.2%, and 9.1% to 19.3%, respectively. © 2022 IEEE.