Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 20 de 151
Filter
1.
Atmosphere ; 13(7):1023, 2022.
Article in English | ProQuest Central | ID: covidwho-1963692

ABSTRACT

(1) Background: To better carry out air pollution control and to assist in accurate investigations of air pollution, in this study, we fully explore the spatial distribution characteristics of air pollution complaint results and provide guidance for air pollution control by combining regional air monitoring data. (2) Methods: By selecting the air pollution complaint information in Beijing from 2019 to 2020, in this study, we extract the names and addresses of complaint points, as well as the complaint times and types by adopting the BERT (bidirectional encoder representations from transformers) + CRF (conditional random field) model deep learning method. Moreover, through further filtering and processing of the complaint points’ address information, we achieve address matching and spatial positioning of the complaint points, and realize the regional spatial representation of air pollution complaints in Beijing in the form of a heat map. (3) Results: The experimental results are compared and analyzed with the ranking data of total suspended particulate (TSP) concentration of townships (streets) in Beijing during the same period, indicating that the key areas of air pollution complaints have a high correlation with the key polluted township (street) areas. The distribution of complaints and the types of complaints in each township (street) differ according to the population density in each township (street), the level of education, and economic activity. (4) Conclusions: The results of this study show that the public, as the intuitive perceiver of air pollution, is sensitive to the air pollution situation at a smaller spatial scale;furthermore, complaints can provide guidance and reference for the direction of air pollution control and law enforcement investigations when coupled with geographical features and economic status.

2.
Agronomy ; 12(7):1583, 2022.
Article in English | ProQuest Central | ID: covidwho-1963665

ABSTRACT

Timely, accurate, and repeatable crop mapping is vital for food security. Rice is one of the important food crops. Efficient and timely rice mapping would provide critical support for rice yield and production prediction as well as food security. The development of remote sensing (RS) satellite monitoring technology provides an opportunity for agricultural modernization applications and has become an important method to extract rice. This paper evaluated how a semantic segmentation model U-net that used time series Landsat images and Cropland Data Layer (CDL) performed when applied to extractions of paddy rice in Arkansas. Classifiers were trained based on time series images from 2017–2019, then were transferred to corresponding images in 2020 to obtain resultant maps. The extraction outputs were compared to those produced by Random Forest (RF). The results showed that U-net outperformed RF in most scenarios. The best scenario was when the time resolution of the data composite was fourteen day. The band combination including red band, near-infrared band, and Swir-1 band showed notably better performance than the six widely used bands for extracting rice. This study found a relatively high overall accuracy of 0.92 for extracting rice with training samples including five years from 2015 to 2019. Finally, we generated dynamic maps of rice in 2020. Rice could be identified in the heading stage (two months before maturing) with an overall accuracy of 0.86 on July 23. Accuracy gradually increased with the date of the mapping date. On September 17, overall accuracy was 0.92. There was a significant linear relationship (slope = 0.9, r2 = 0.75) between the mapped areas on July 23 and those from the statistical reports. Dynamic mapping is not only essential to assist farms and governments for growth monitoring and production assessment in the growing season, but also to support mitigation and disaster response strategies in the different growth stages of rice.

3.
Journal of Sensors ; 2022, 2022.
Article in English | ProQuest Central | ID: covidwho-1962466

ABSTRACT

The recognition of aircraft wake vortex can provide an indicator of early warning for civil aviation transportation safety. In this paper, several wake vortex recognition models based on deep learning and traditional machine learning were presented. Nonetheless, these models are not completely suitable owing to their dependence on the visualization of LiDAR data that yields the information loss of in reconstructing the behavior patterns of wake vortex. To tackle this problem, we proposed a lightweight deep learning framework to recognize aircraft wake vortex in the wind field of Shenzhen Baoan Airport’s arrival and departure routes. The nature of the introduced model is geared towards three aspects. First, the dilation patch embedding module is used as the input representation of the framework, attaining additional rich semantics information over long distances while maintaining parameters. Second, we combined a separable convolution module with a hybrid attention mechanism, increasing the model’s attention to the space position of wake vortex core. Third, environmental factors that affect the vortex behavior of the aircraft’s wake were encoded into the model. Experiments were conducted on a Doppler LiDAR acquisition dataset to validate the effectiveness of the proposed model. The results show that the proposed network has an accuracy of 0.9963 and a recognition speed at 100 frames per second was achieved on an experimental device with 0.51 M parameters.

4.
15th International Conference on PErvasive Technologies Related to Assistive Environments, PETRA 2022 ; : 615-621, 2022.
Article in English | Scopus | ID: covidwho-1962418

ABSTRACT

Recent studies indicate that detecting radiographic patterns on CT scans can yield high sensitivity and specificity for Covid-19 localization. In this paper, we investigate the appropriateness of deep learning models transferability, for semantic segmentation of pneumonia-infected areas in CT images. Transfer learning allows for the fast initialization/reutilization of detection models, given that large volumes of training data are not available. Our work explores the efficacy of using pre-trained U-Net architectures, on a specific CT data set, for identifying Covid-19 side-effects over images from different datasets. Experimental results indicate improvement in the segmentation accuracy of identifying Covid-19 infected regions. © 2022 ACM.

5.
2nd International Conference on Technology Enhanced Learning in Higher Education, TELE 2022 ; : 343-347, 2022.
Article in English | Scopus | ID: covidwho-1961428

ABSTRACT

The article prepared for publication contains material on digitalization that is an integral part of the life of modern society, providing acceleration of economic processes and saving time and resources in almost all sectors of the economy, including education as the fundamental basis of human development. The purpose of the study is to provide in-depth justification of the prospects of digital technologies in the educational process for its qualitative improvement and subsequent growth of labor productivity and strengthening of national economic potential. Its achievement is focused on revealing the place and role of digitalization for education and substantiating the needs of relevant institutions in it with an emphasis on the pandemic period. The solution of the formulated tasks is carried out by using monographic and dialectical methods that help to systematize and improve the accumulated knowledge about education and the scientific and technical achievements used in it on the basis of the identified advantages and interpretation of their semantic content. These advantages are driving factors in the development of educational processes. The material presented in the article is useful for managers, specialists involved in the development of educational processes, as well as teachers of pedagogical directions. © 2022 IEEE.

6.
2022 Workshop on Scientific Document Understanding, SDU 2022 ; 3164, 2022.
Article in English | Scopus | ID: covidwho-1958163

ABSTRACT

MeSH (Medical Subject Headings) is a large thesaurus created by the National Library of Medicine and used for fine-grained indexing of publications in the biomedical domain. In the context of the COVID-19 pandemic, MeSH descriptors have emerged in relation to articles published on the corresponding topic. Zero-shot classification is an adequate response for timely labeling of the stream of papers with MeSH categories. In this work, we hypothesise that rich semantic information available in MeSH has potential to improve BioBERT representations and make them more suitable for zero-shot/few-shot tasks. We frame the problem as determining if MeSH term definitions, concatenated with paper s are valid instances or not, and leverage multi-task learning to induce the MeSH hierarchy in the representations thanks to a seq2seq task. Results establish a baseline on the MedLine and LitCovid datasets, and probing shows that the resulting representations convey the hierarchical relations present in MeSH. © 2021 Copyright for this paper by its authors.

7.
IEEE Transactions on Knowledge and Data Engineering ; : 1-14, 2022.
Article in English | Scopus | ID: covidwho-1948851

ABSTRACT

The wide spread of fake news is increasingly threatening both individuals and society. Great efforts have been made for automatic fake news detection on a single domain (e.g., politics). However, correlations exist commonly across multiple news domains, and thus it is promising to simultaneously detect fake news of multiple domains. Based on our analysis, we pose two challenges in multi-domain fake news detection: 1) <bold/>domain shift<bold/>, caused by the discrepancy among domains in terms of words, emotions, styles, etc. 2) <bold/>domain labeling incompleteness<bold/>, stemming from the real-world categorization that only outputs one single domain label, regardless of topic diversity of a news piece. In this paper, we propose a Memory-guided Multi-view Multi-domain Fake News Detection Framework (M <inline-formula><tex-math notation="LaTeX">$

8.
23rd International Symposium on Quality Electronic Design, ISQED 2022 ; 2022-April, 2022.
Article in English | Scopus | ID: covidwho-1948806

ABSTRACT

COVID-19 (Corona Virus Disease 2019) is a pandemic which has been spreading exponentially around the globe. Many countries had adopted stay-at-home or lockdown policies to control its spreading. However, prolonged stay-at-home can cause worse effects like economic crises, unemployment, food scarcity, and mental health problems of individuals. EasyBand2.0 is a wearable personal safety device that helps in social distancing and also helps in safe mobility. Under the IoMT (Internet of Medical Things) framework the wearable EasyBand2.0 device helps in social distancing, it avoids human-to-human contact and helps maintain a safer distance. EasyBand2.0 uses the Low Power BLE technology to sense distance between two user devices and alert them based on the distance and time spent in proximity. Safe mobility of people is also important as travel is resumed in all forms. This paper proposes a software application along with the easy band to further be integrated with a system that works based on GPS (Global Positioning System) or GIS (Geographic Information System) to provide travel logging for contact tracing without exposing personal data. A CARS (Context Aware Recommendation System) based safe zone recommender system is proposed in this paper to aid safe mobility. © 2022 IEEE.

9.
International Journal of Advanced Computer Science and Applications ; 13(6):337-345, 2022.
Article in English | Scopus | ID: covidwho-1934696

ABSTRACT

This study used opinion mining theory and the potentials of artificial intelligence to explore the opinions, sentiments, and attitudes of customers expressed on Twitter regarding the services provided by the Saudi telecommunications companies during the COVID-19 crisis. A corpus of 12,458 Twitter posts was constructed covering the period 2020–2021. For data analysis, the study adopted a discourse-based mining approach, combining vector space classification (VSC) and collocation analysis. The results indicate that most users had negative attitudes and sentiments regarding the performance of the telecommunications companies during the pandemic, as reflected in both the lexical semantic properties and discoursal and thematic features of their Twitter posts. The study of collocates and the discoursal properties of the data was useful in attaining a deeper understanding of the users’ responses and attitudes to the performance of the telecommunications companies during the COVID-19 pandemic. It was not possible for text clustering based on the “bag of words” model alone to address the discoursal features in the corpus. Opinion mining applications, especially in Arabic, thus need to integrate discourse approaches to gain a better understanding of people’s opinions and attitudes regarding given issues © 2022. International Journal of Advanced Computer Science and Applications.All Rights Reserved.

10.
Applied Sciences ; 12(13):6615, 2022.
Article in English | ProQuest Central | ID: covidwho-1933961

ABSTRACT

Featured ApplicationAuthors are encouraged to provide a concise description of the specific application or a potential application of the work. This section is not mandatory.The research aim is to construct a disease-symptom knowledge graph (DSKG) as a cause-effect knowledge graph containing disease-symptom relations as a cause-effect relation type determined from downloaded documents on medical web-board resources. Each disease-symptom relation connects a disease-name concept node (a causative-concept node) to a corresponding node having a group of correlated symptom-concept/effect-concept features as common symptom-concept/effect-concept features among some disease-name concepts. The DSKG benefits non-professionals in preliminary diagnosis through a recommender web-board. There are three main problems: how to determine symptom concepts from sentences without annotation on the documents having disease-name concepts as the documents’ topic-names;how to determine the disease-symptom relations from the documents with/without complications;and how to construct the DSKG involving high dimensional symptom-concept features after union of the correlated symptom-concept groups. Therefore, we apply a word co-occurrence pattern including medical-symptom expressions from Wikipedia including MeSH and the Lexitron Dictionary to determine the symptom concepts. The Cartesian product is applied for automatic-supervised machine learning to determine the disease-symptom relation. We propose using Principal Component Analysis for constructing the DSKG by dimensionality reduction in the symptom-concept features with minimized information loss. In contrast to previous works, the proposed approach enables the DSKG construction with precise and concise representation scores of 7.8 and 9, respectively.

11.
2022 International Conference on Optics and Machine Vision, ICOMV 2022 ; 12173, 2022.
Article in English | Scopus | ID: covidwho-1932600

ABSTRACT

Covid-19 pandemic continues to threat health of the global population, an efficient way to restrain the Covid-19 outbreak is timely screening suspected cases for quarantine and treatment. Despite of pathogenic laboratory testing is the gold standard to screen suspected cases, but it may obtain false negative results and consuming a lot of time. Computed tomography of chest can be an alternative diagnostic method to screen suspected cases that is based on radio graphical changes in lung area of Covid-19 confirmed case. Precisely delineate the lung area that is first and critical step for screening computed tomography image of chest by using deep learning method. In this paper, several related previous works will be introduced at first. Then, an improved encoder-decoder based segmentation framework is proposed, which is integrated with multi-scale densely connection-based convolution block and skip connection path. Moreover, in model training process, a semi-supervised manner is applied to train model which can reduce the demand of labeled training data. Finally, the proposed method is tested with a dataset of public X-ray image of chest. The experiment test proposed model in this paper with varieties of segmentation methods and result demonstrates promising performance of proposed model that against several other deep learning models. © 2022 SPIE

12.
Journal for the Study of Religions and Ideologies ; 21(62):70-83, 2022.
Article in English | ProQuest Central | ID: covidwho-1929197

ABSTRACT

This research investigates the kinds of speech acts contained in syarah Al-Hikam. In this case, syarah is defined as one form of explanation of a matan (substance) from Al-Hikam aphorism (the book of wisdoms) in Islamic mysticism as the base foundation of spiritual counselling. The language of syarah has a distinctive character, such as: the process of explaining to the reader what is being described about matan. Therefore, it is important to study language of syarah based on the pragmatics aspects, especially the variants of speech acts contained in syarah al-Hikam as a set of values and principles that strengthen the relationship between the speaker and the reader and improve the Islamic healing process. In addition, this research also elaborated how the speech acts are translated from Arabic (Ll) into Bahasa Indonesia (L2). Based on the explanation, this study examines several problems, such as: (l) what are the forms of speech acts contained in syarah al-Hikam, (2) what translation techniques are used by the translator in translating speech acts contained in syarah al-Hikam from Source Language (SL) to Target Language (TL), and (3) how is the relation between the translation quality of syarah al-Hikam speech acts and the spiritual counselling. The method in this research was divided into three stages. The first stage dealt with data collection method. The data were collected through observation and classification of speech acts in syarah al-Hikam. The second stage was about data analysis model. The data were analyzed using Spradley model, namely: (l) domain analysis, (2) taxonomic analysis, (3) a component analysis, and (4) cultural themes analysis. Meanwhile, the third stage was related to the report on the findings in which informal method, in the form of narrations and descriptions of various speech acts in syarah al-Hikam and the pragmatic equivalence of these speech acts, was used. This study does not examine the most dominant form of speech acts, but examining how the character of syarah language in religious books as the base foundation of Islamic spiritual counselling applies the pragmatics principles. Thus, this article examines how a pragmatic equivalence between source language and target language is achieved.

13.
J Korean Acad Nurs ; 52(3): 291-307, 2022 Jun.
Article in Korean | MEDLINE | ID: covidwho-1928740

ABSTRACT

PURPOSE: The aims of study were to identify the main keywords, the network structure, and the main topics of press articles related to nurses that have appeared in media reports. METHODS: Data were media articles related to the topic "nurse" reported in 16 central media within a one-year period spanning July 1, 2019 to June 30, 2020. Data were collected from the Big Kinds database. A total of 7,800 articles were searched, and 1,038 were used for the final analysis. Text network analysis and topic modeling were performed using NetMiner 4.4. RESULTS: The number of media reports related to nurses increased by 3.86 times after the novel coronavirus (COVID-19) outbreak compared to prior. Pre- and post-COVID-19 network characteristics were density 0.002, 0.001; average degree 4.63, 4.92; and average distance 4.25, 4.01, respectively. Four topics were derived before and after the COVID-19 outbreak, respectively. Pre-COVID-19 example topics are "a nurse who committed suicide because she could not withstand the Taewoom at work" and "a nurse as a perpetrator of a newborn abuse case," while post-COVID-19 examples are "a nurse as a victim of COVID-19," "a nurse working with the support of the people," and "a nurse as a top contributor and a warrior to protect from COVID-19." CONCLUSION: Topic modeling shows that topics become more positive after the COVID-19 outbreak. Individual nurses and nursing organizations should continuously monitor and conduct further research on nurses' image.


Subject(s)
COVID-19 , Nurses , Disease Outbreaks , Humans , Infant, Newborn , SARS-CoV-2
14.
Bulletin of Ugric Studies ; 12(1):48-56, 2022.
Article in English | Scopus | ID: covidwho-1924944

ABSTRACT

Introduction: the COVID-19 pandemic has a major impact on the various arenas of communication. The most striking language changes can be observed in the lexicon. Since 2020, thousands of new pandemic-related words and phrases, so-called virologisms, appeared in many languages of the world, including the Hungarian, English, German and Russian. It is worthwhile to examine virologisms from a word-formation aspect and a semantic aspect. The present study focuses mainly on the semantics of the Hungarian virologisms. Linguistic examples are reviewed from the aspect of lexical (i.e. word) relationships such as synonymy, polysemy, or homonymy. Objective: to identify the semantics of virologisms of the Hungarian language (in comparison with the English, German and Russian). Research materials: published articles, studies and dictionaries in the Hungarian, Russian and English languages with particular regard to the Karanténszótár, COVIDictionary and Slovarj russkogo jazyka koronavirusnoj epochi (Dictionary of the Russian Language of the Coronavirus Era). Results and novelty of the research: to the best of the authors’ knowledge, no comprehensive work presenting an analysis of virologisms from a semantic aspect has been published, up to the beginning of 2022 in the Hungarian or in any other language they are familiar with, as the majority of the studies approach the problem from a formal perspective. The present study points out intra-lexical and inter-lexical semantic relationships (e.g. synonymy, polysemy, and homonymy) that can be observed between lexemes belonging to the same language register and those belonging to different language registers: while synonymy makes the talk about the pandemic more colorful, polysemy and homonymy can sometimes be a source of misunderstanding. © 2022 Ob-Ugric Institute of Applied Researches and Development. All rights reserved.

15.
Theory and Practice in Language Studies ; 12(7):1286-1293, 2022.
Article in English | ProQuest Central | ID: covidwho-1924778

ABSTRACT

Metaphors permeate our daily communication, and they are part of our cognition. The present study investigates metaphors in a corpus-based study during the Coronavirus disease 19 crisis (COVID-19) using the Antconc Software. The way written media discourse framed the COVID-19 Crisis, especially in the Middle East received little attention from discourse analysts. The data include news editorials about the Coronavirus disease 19 from April 1, 2020 to July 5, 2020 collected from "the Jordan Times" Newspaper in English. The metaphors will be analysed according to Lakoff and Johnson's (1980;2003) perspective of Conceptual Metaphors and Charteris-Black (2004) of Critical Metaphor Analysis. The quantitative analysis shows that the conceptual metaphors COVID-19 IS WAR, COVID-19 IS WATER, and COVID-19 IS A PERSON are highly used in the corpus to frame the pandemic. The conceptual metaphor COVID-19 IS WAR is not only used to represent a war against the disease, but also a war between countries. So, the metaphorical use is politicised, and reflects hidden ideology. The quantitative analysis asserts that the context is the decisive factor for the analysis of certain lexical items related to the pandemic and identifying whether they are literally or metaphorically used.

16.
Medical Imaging 2022: Computer-Aided Diagnosis ; 12033, 2022.
Article in English | Scopus | ID: covidwho-1923076

ABSTRACT

Automated analysis of chest imaging in coronavirus disease (COVID-19) has mostly been performed on smaller datasets leading to overfitting and poor generalizability. Training of deep neural networks on large datasets requires data labels. This is not always available and can be expensive to obtain. Self-supervision is being increasingly used in various medical imaging tasks to leverage large amount of unlabeled data during pretraining. Our proposed approach pretrains a vision transformer to perform two self-supervision tasks - image reconstruction and contrastive learning on a Chest Xray (CXR) dataset. In the process, we generate more robust image embeddings. The reconstruction module models visual semantics within the lung fields by reconstructing the input image through a mechanism which mimics denoising and autoencoding. On the other hand, the constrastive learning module learns the concept of similarity between two texture representations. After pretraining, the vision transformer is used as a feature extractor towards a clinical outcome prediction task on our target dataset. The pretraining multi-kaggle dataset comprises 27499 CXR scans while our target dataset contains 530 images. Specifically, our framework predicts ventilation and mortality outcomes for COVID-19 positive patients using baseline CXR. We compare our method against a baseline approach using pretrained ResNet50 features. Experimental results demonstrate that our proposed approach outperforms the supervised method. © 2022 SPIE.

17.
7th International Conference on Image and Signal Processing and their Applications, ISPA 2022 ; 2022.
Article in English | Scopus | ID: covidwho-1922720

ABSTRACT

Deep learning (DL) algorithms are widely applied in many disciplines such as medical imaging, bioinformatics, and computer vision. DL models have been used in medical imaging to perform image segmentation, classification, and detection. During the outbreak of the COVID-19 pandemic, DL has been extensively used to develop COVID-19 screening systems. The reverse transcription-polymerase chain reaction (RT-PCR) is the gold standard method for COVID-19 screening. However, DL has been proposed to detect patients infected with COVID-19 through radiological imaging in Chest X-rays and chest computed tomography (CT) images. This paper proposes transfer learning to train modified U-Net models to segment the COVID-19 chest CT images into two regions of lung infection (ground-glass and consolidation). The proposed modified U-Net models were constructed by replacing the encoder part with a pre-trained convolutional neural network (CNN) model. Three pre-trained CNN models, namely, EfficientNet-b0, EfficientNet-b1, and EfficientNet-b2 were used. The proposed models were evaluated on the COVID-19 CT Images Segmentation dataset available in an open Kaggle challenge. The obtained results show that the proposed EfficientNet-b2_U-Net model yielded the highest FScore of 0.5666. © 2022 IEEE.

18.
6th International Conference on Intelligent Computing and Control Systems, ICICCS 2022 ; : 1399-1403, 2022.
Article in English | Scopus | ID: covidwho-1922680

ABSTRACT

An outbreak of the coronavirus disease due to occur in 2020 has already had a significant impact on the human race. Wearing a "Face Mask"and maintaining "Social Distancing"are the only means to protect ourselves from this pandemic. Several service providers, such as airlines, hotels, hospitals, and train stations, demand their customers to access the service only if the Mask is worn correctly and social distance is maintained. Manually checking to see if the rule of mask wearing and social distance is being observed is impossible due to the significant human resource consumption. As a one-stage detector, the COVID-19 Face Mask and Social Distancing Detector System uses an artificial neural network to combine high-level semantic information with various feature maps and a machine learning module to identify face masks and social distances at the same time. It will also be able to detect persons without masks and the violence of social separation by using existing IP cameras, CCTV cameras, and computer vision. This technology eliminates the need for a manual surveillance system by providing instruments for safety and security. The technology can be used in any type of infrastructure, including hospitals, government offices, schools, and construction sites. Therefore, the face mask and social distance detector system developed, could aid to secure the protection and security of ourselves and our loved ones. © 2022 IEEE.

19.
Slavia Centralis ; 15(1):327-346, 2022.
Article in Russian, Slovenian | Scopus | ID: covidwho-1918596

ABSTRACT

The purpose of this research is to study new vocabulary and phraseology that appeared in relation to the coronavirus pandemic in different groups of Slavic languages: Ukrainian, Russian and Czech. Thematic groups, figurative and word-shaping resources are described. The results of the research revealed parallel trends: Mainly borrowings from the English language and occasional neologisms with a clear expressive tone. Innovations in the fields of Phraseology and Paremiology take place based on the existing structural-semantic models and transformations of proverbs into new anti-proverbs. © 2022 University of Maribor. All rights reserved.

20.
Ieee Access ; 10:66757-66769, 2022.
Article in English | Web of Science | ID: covidwho-1915929

ABSTRACT

Image inpainting techniques have been greatly improved by relying on structure and texture priors. However, damaged original images or rough predictions cannot provide sufficient texture information and accurate structural priors, leading to a drop in image quality. Moreover, from the perspective of human visual perception, it is important to pay attention to facial symmetry and facial attribute consistency. In this paper, we present a face inpainting system with iteration structure, guided by generative facial priors contained in pretrained GANs and predicted semantic information. Specifically, generative facial priors generated by the GAN inversion techniques introduce sufficient textures and features to assist inpainting;semantic maps are able to provide facial structural information and semantic categories of different pixels for face reconstruction. In particular, we iteratively refine images multiple times, updating semantic maps at each iteration. The Weighted Prior-Guidance Modulation layer (WPGM) is devised for incorporating priors into networks through spatial modulation. We also propose facial feature self-symmetry loss to constrain the symmetry of faces in feature space. Experiments on CelebA-HQ and LaPa datasets demonstrate the superiority of our model for facial detail and attribute consistency. Meanwhile, under the background of COVID-19, it is worth trying recognition via inpainting to deal with recognition challenges brought by mask occlusion. Relevant experiments show that our inpainting model does help to recognition tasks to a certain degree, with higher accuracy.

SELECTION OF CITATIONS
SEARCH DETAIL