Required packages
以下が、このプログラムの実行に必要なパッケージ
require(easyPubMed)## Loading required package: easyPubMedrequire(tm)## Loading required package: tm## Loading required package: NLPrequire(udpipe)## Loading required package: udpiperequire(wordcloud)## Loading required package: wordcloud## Loading required package: RColorBrewerrequire(word2vec)## Loading required package: word2vecrequire(Rtsne)## Loading required package: Rtsnerequire(plotly)## Loading required package: plotly## Loading required package: ggplot2##
## Attaching package: 'ggplot2'## The following object is masked from 'package:NLP':
##
## annotate##
## Attaching package: 'plotly'## The following object is masked from 'package:ggplot2':
##
## last_plot## The following object is masked from 'package:stats':
##
## filter## The following object is masked from 'package:graphics':
##
## layoutData preparation
PubMedからデータを取得する
query <- "COVID 19"
ids <- get_pubmed_ids(query)
ids$Count## [1] "324847"324,798件がヒットする（2022.12.26 現在）
内容を減らすために、クエリを限定的にする。
query <- "COVID 19 AND machine learning"
ids <- get_pubmed_ids(query)
ids$Count## [1] "4372"4,372件がヒットする（2022.12.26 現在）
これを使って以下の解析を行う。
まず、PubMedのデータを取り出してくる。fetch_pubmed_data関数は、1度に最大4999件まで取り出せる。 （それ以上取り出す場合は、開始番号を変えながら、繰り返し関数を使う必要がある）。
pmd.xml <- fetch_pubmed_data(ids, retmax = 5000)
pmd.list <- articles_to_list(pmd.xml)
length(pmd.list)## [1] 4372タイトルと要旨（アブストラクト）からなるデータフレームを作成する。また、欠測がある（要旨が無い）論文を除いておく。
titl <- rep(NA, length(pmd.list))
abst <- rep(NA, length(pmd.list))
for(i in 1:length(pmd.list)) {
df <- article_to_df(pmd.list[[i]], max_chars = -1, getAuthors = F)
titl[i] <- df$title
abst[i] <- df$abstract
}
df <- data.frame(titl, abst)
df <- na.omit(df)
dim(df)## [1] 4275 2Analysis of the titles
タイトルに現れる単語の出現頻度を調べる。
オリジナルのタイトルのデータを確認。
doc <- df$titl
doc[1:2]## [1] "New advances in prediction and surveillance of preeclampsia: role of machine learning approaches and remote monitoring."
## [2] "Comparing Short-Term Univariate and Multivariate Time-Series Forecasting Models in Infectious Disease Outbreak."全て小文字に変換し、数字や、カッコや句読点を取り除く。
その前に、ハイフンをスペースに変換しておく。
doc.tmp <- gsub("-", " ", doc)
doc[1:2]## [1] "New advances in prediction and surveillance of preeclampsia: role of machine learning approaches and remote monitoring."
## [2] "Comparing Short-Term Univariate and Multivariate Time-Series Forecasting Models in Infectious Disease Outbreak."doc.tmp[1:2]## [1] "New advances in prediction and surveillance of preeclampsia: role of machine learning approaches and remote monitoring."
## [2] "Comparing Short Term Univariate and Multivariate Time Series Forecasting Models in Infectious Disease Outbreak."小文字に変換し、数字や、カッコや句読点を取り除く。
doc.cleaned <- stripWhitespace(
removePunctuation(
removeNumbers(tolower(doc))))
doc.cleaned[1:2]## [1] "new advances in prediction and surveillance of preeclampsia role of machine learning approaches and remote monitoring"
## [2] "comparing shortterm univariate and multivariate timeseries forecasting models in infectious disease outbreak"頻度をカウントする（論文ごと）。
dtf <- document_term_frequencies(doc.cleaned)
head(dtf, 20)## doc_id term freq
## 1: doc1 new 1
## 2: doc1 advances 1
## 3: doc1 in 1
## 4: doc1 prediction 1
## 5: doc1 and 2
## 6: doc1 surveillance 1
## 7: doc1 of 2
## 8: doc1 preeclampsia 1
## 9: doc1 role 1
## 10: doc1 machine 1
## 11: doc1 learning 1
## 12: doc1 approaches 1
## 13: doc1 remote 1
## 14: doc1 monitoring 1
## 15: doc2 comparing 1
## 16: doc2 shortterm 1
## 17: doc2 univariate 1
## 18: doc2 and 1
## 19: doc2 multivariate 1
## 20: doc2 timeseries 1全ての論文に対して頻度を足し合わせる。上位100位の単語を示す。
res <- tapply(dtf$freq, dtf$term, sum)
sort(res, decreasing = T)[1:100]## of covid and learning the
## 2959 2830 2268 1645 1575
## for in a machine using
## 1567 1462 1425 1090 949
## to on deep with patients
## 732 694 639 609 451
## prediction from detection sarscov pandemic
## 445 420 398 396 387
## analysis study model based data
## 382 361 360 338 328
## an approach chest images during
## 314 311 295 282 272
## artificial diagnosis intelligence disease ct
## 256 253 227 216 215
## health classification clinical xray by
## 208 206 205 204 201
## models predicting coronavirus novel review
## 197 188 183 172 171
## risk infection mortality severity system
## 168 163 163 156 151
## development network pneumonia neural learningbased
## 148 147 144 142 128
## predict social early forecasting features
## 128 123 109 105 102
## framework application impact networks drug
## 102 100 99 98 96
## lung factors validation techniques screening
## 96 95 95 94 93
## care methods approaches case ensemble
## 91 91 88 85 85
## modeling predictive algorithm identification image
## 85 85 84 84 84
## automated method vaccine medical transfer
## 83 83 83 81 81
## public among against cases through
## 80 79 78 77 77
## human feature new twitter algorithms
## 75 73 73 73 72
## digital media potential respiratory as
## 71 71 71 70 69よくある（あまり意味をもたない）単語を取り除く。そのための単語のリストを準備する。
stp <- stopwords("en")
head(stp, 20)## [1] "i" "me" "my" "myself" "we"
## [6] "our" "ours" "ourselves" "you" "your"
## [11] "yours" "yourself" "yourselves" "he" "him"
## [16] "his" "himself" "she" "her" "hers"上のリストのいずれかに一致する場合はデータから除く。
selector <- !(dtf$term %in% stopwords())
dtf.sel <- dtf[selector, ]再度、数え上げをする。上位100位を示す。
word.count <- tapply(dtf.sel$freq, dtf.sel$term, sum)
sort(word.count, decreasing = T)[1:100]## covid learning machine using deep
## 2830 1645 1090 949 639
## patients prediction detection sarscov pandemic
## 451 445 398 396 387
## analysis study model based data
## 382 361 360 338 328
## approach chest images artificial diagnosis
## 311 295 282 256 253
## intelligence disease ct health classification
## 227 216 215 208 206
## clinical xray models predicting coronavirus
## 205 204 197 188 183
## novel review risk infection mortality
## 172 171 168 163 163
## severity system development network pneumonia
## 156 151 148 147 144
## neural learningbased predict social early
## 142 128 128 123 109
## forecasting features framework application impact
## 105 102 102 100 99
## networks drug lung factors validation
## 98 96 96 95 95
## techniques screening care methods approaches
## 94 93 91 91 88
## case ensemble modeling predictive algorithm
## 85 85 85 85 84
## identification image automated method vaccine
## 84 84 83 83 83
## medical transfer public among cases
## 81 81 80 79 77
## human feature new twitter algorithms
## 75 73 73 73 72
## digital media potential respiratory monitoring
## 71 71 71 70 67
## cohort hybrid automatic evaluation convolutional
## 66 66 65 65 64
## assessment blood healthcare hospitalized role
## 62 62 62 60 60
## sentiment outcomes systematic identifying imaging
## 60 59 58 57 57ワードクラウドを用いて表示する。頻出単語上位100を表示する。
word.top100 <- sort(word.count, decreasing = T)[1:100]
wordcloud(names(word.top100), freq = word.top100, color = brewer.pal(8, "Dark2"))
Analysis with word2vec
Word2vecを使った解析を行う。なお、Word2vecについての説明は、原著https://arxiv.org/pdf/1301.3781.pdf、また、解説https://arxiv.org/pdf/1411.2738.pdfの論文を参考にするとよい（後者が分かりやすい）。なお、ブログなどの記事も多い。例えば、https://israelg99.github.io/2017-03-23-Word2Vec-Explained/、https://towardsdatascience.com/word2vec-explained-49c52b4ccb71。
まずは、データを準備する。
x <- txt_clean_word2vec(df$abst)次に、word2vec関数で、単語間の関係を学習する。ここでは、skip-gramアルゴリズムを用いる。
model <- word2vec(x, type = "skip-gram", dim = 30, window = 5, iter = 10)結果を表示。単語がベクトル空間内の点として表される。
head(as.matrix(model), 10)## [,1] [,2] [,3] [,4] [,5]
## access 1.8834980 -2.3609595 1.13500381 1.4927373 0.542817116
## stability -0.8966581 -1.0324870 -0.31039262 -0.5030487 0.038705207
## eliminating 1.5642663 -0.6669895 0.34348664 0.3617151 1.166107416
## ligands 1.6098195 -0.4321874 0.60072803 0.1811531 0.224953234
## represent 0.3369185 -1.8443726 0.04610289 0.6622514 -1.278716207
## spanning 0.7340813 0.8242792 1.36835647 0.6356338 -0.723508716
## outpatient 1.7866665 -0.2589226 1.98418868 1.2449813 0.883127272
## degeneration 0.5019498 0.3133734 1.40740240 0.9938579 -0.002220026
## unparalleled 0.8866035 0.6266905 0.93503529 0.8410850 0.064271405
## insights 1.3087177 -1.1826060 -0.55142534 0.5043691 -0.693867624
## [,6] [,7] [,8] [,9] [,10]
## access -0.02805615 -1.55603015 0.38960561 0.41539526 -0.45791730
## stability -1.29562759 -0.26410103 -0.48054999 1.57208276 -0.20944515
## eliminating -0.02478027 -1.10254073 0.07579076 1.66908717 0.52693337
## ligands -0.64256972 0.03663487 0.06771453 0.46551913 -0.86478013
## represent 0.38264179 -1.02044368 1.46145773 1.09471750 -2.01458192
## spanning -0.42140484 1.35439456 1.62295806 0.04233977 -0.16239795
## outpatient 0.89739275 1.31969237 0.08918275 1.53049541 -0.04111318
## degeneration 1.73076141 -0.23754048 -1.40735209 -0.02257124 -2.45811152
## unparalleled -0.59475613 -1.25391376 1.05681419 0.48143473 -1.14191377
## insights -0.42294213 -0.44819540 0.36971584 -0.29888809 -1.18783474
## [,11] [,12] [,13] [,14] [,15]
## access -1.5041744 -1.0554111 -1.08506060 -0.2090475 -0.42590329
## stability -1.1446419 -0.6541144 -0.46916291 1.6136422 1.30108607
## eliminating -0.2792227 -0.1273677 1.76837897 2.1591556 -0.17860299
## ligands -0.2024247 1.4765788 -1.02697039 1.8706492 1.76170230
## represent 0.1696163 0.2273275 0.10218593 0.9292352 -0.02506509
## spanning 0.6705174 -0.8349549 -0.13178769 -0.1140475 1.80735731
## outpatient -0.7826992 -1.1549877 -0.22847484 -0.9016120 -0.75434852
## degeneration -1.1211474 -0.4919341 -0.39300355 0.2923086 -1.01190710
## unparalleled -1.6372910 -1.6781262 -0.09766967 -0.6491150 0.11443575
## insights 0.3033503 -0.6893293 1.39802897 -0.1052159 0.05170381
## [,16] [,17] [,18] [,19] [,20] [,21]
## access -0.4937030 -0.35358435 0.1562183 0.34102646 1.06768548 1.4833052
## stability 0.8501904 1.03535473 -1.5515784 2.87455153 0.31837282 0.3806680
## eliminating 0.1075667 0.27414146 -1.6702029 0.93148786 0.70044762 1.5865083
## ligands 0.8309996 0.36098820 -0.8173749 2.23008323 0.69616765 0.8623462
## represent -1.2056096 -0.62799406 -1.7664572 1.03400826 -0.41982606 0.5974448
## spanning -0.8194700 0.40543330 1.1067102 0.85468942 0.12250974 0.9896046
## outpatient 1.1913366 -0.01630457 0.7075865 0.02031492 0.69213843 0.8853406
## degeneration -1.3654140 -0.10030320 0.4742510 0.78608686 -0.22139725 1.4619242
## unparalleled 2.0729208 -0.33632502 -0.7768776 0.71200866 0.05488064 1.3273437
## insights 0.5510238 0.02700333 -1.0044469 1.57278991 0.04942456 2.7856309
## [,22] [,23] [,24] [,25] [,26]
## access 1.05134892 -0.3543068 0.1574571 -0.7036515 0.02190759
## stability -1.50188696 -0.6507681 -0.5047027 0.2460931 0.10925783
## eliminating -1.23322070 -0.7997144 0.1323705 0.7459010 0.04732146
## ligands -0.66868216 -0.4972401 -2.0995812 0.5459074 0.39897972
## represent 0.06937215 -1.2207992 -0.3988610 0.8187262 0.03318119
## spanning -0.95820498 1.1151983 -0.2620923 -1.9664277 0.46002764
## outpatient -0.19677567 1.1265121 1.9988035 -1.3363581 -0.02731497
## degeneration -0.36374572 -1.3789468 1.4970579 0.3939926 -0.88063055
## unparalleled 1.77237880 -0.4483878 1.0660981 -0.4135475 -0.58891791
## insights 0.68586522 -1.1481521 1.1800277 1.0301284 -1.45251751
## [,27] [,28] [,29] [,30]
## access 1.18207002 0.90332258 -1.23154032 0.03696854
## stability 0.35127428 0.28652859 -0.09005820 1.04790318
## eliminating 0.03140764 -0.75984484 0.01902853 1.79779601
## ligands -0.44234350 0.03135437 0.95867807 1.07195950
## represent 1.52137721 0.11871677 -1.88761401 0.29222265
## spanning 2.20786238 -0.84908479 -0.37805405 1.01536453
## outpatient 0.53431118 -1.04972351 -0.26777437 -0.67968041
## degeneration 1.05353963 -0.52781159 -0.87775630 0.75818646
## unparalleled 0.13198513 1.00854647 -1.83406508 0.18777904
## insights 1.65858340 0.02616262 0.43860298 -0.67190403“mask”という単語と類似度が高い単語を30個リストアップする。
nn <- predict(model, c("mask"), type = "nearest", top_n = 30)
nn## $mask
## term1 term2 similarity rank
## 1 mask wearing 0.9168410 1
## 2 mask masks 0.8957075 2
## 3 mask improper 0.8813931 3
## 4 mask face 0.8736871 4
## 5 mask ddos 0.8611960 5
## 6 mask mandates 0.8553889 6
## 7 mask violations 0.8536556 7
## 8 mask mandate 0.8481603 8
## 9 mask distancing 0.8362848 9
## 10 mask wear 0.8217394 10
## 11 mask masked 0.8132742 11
## 12 mask passengers 0.8108311 12
## 13 mask facemask 0.8093920 13
## 14 mask stores 0.8080394 14
## 15 mask recommending 0.8063456 15
## 16 mask universal 0.8052245 16
## 17 mask hands 0.8020772 17
## 18 mask usage 0.8015509 18
## 19 mask vehicle 0.7990443 19
## 20 mask norms 0.7957740 20
## 21 mask distance 0.7943754 21
## 22 mask effectiveness 0.7909495 22
## 23 mask closure 0.7897944 23
## 24 mask slow 0.7890968 24
## 25 mask perfect 0.7890829 25
## 26 mask overfitting 0.7890803 26
## 27 mask touch 0.7884932 27
## 28 mask respirators 0.7874386 28
## 29 mask fake 0.7872146 29
## 30 mask bans 0.7866435 30“omicron”という単語と類似度が高い単語を30個リストアップする。
nn <- predict(model, c("omicron"), type = "nearest", top_n = 30)
nn## $omicron
## term1 term2 similarity rank
## 1 omicron variant 0.9722302 1
## 2 omicron variants 0.9365270 2
## 3 omicron delta 0.9097216 3
## 4 omicron strains 0.8974876 4
## 5 omicron emergence 0.8811155 5
## 6 omicron alpha 0.8700840 6
## 7 omicron voc 0.8692313 7
## 8 omicron ancestral 0.8593056 8
## 9 omicron mutated 0.8506261 9
## 10 omicron zika 0.8488913 10
## 11 omicron epidemic 0.8485928 11
## 12 omicron causative 0.8456267 12
## 13 omicron zoonotic 0.8451722 13
## 14 omicron outbreaks 0.8441400 14
## 15 omicron emerging 0.8428910 15
## 16 omicron mutant 0.8427690 16
## 17 omicron outburst 0.8424161 17
## 18 omicron beta 0.8403398 18
## 19 omicron lineages 0.8382600 19
## 20 omicron evolves 0.8371908 20
## 21 omicron virus 0.8370464 21
## 22 omicron hotspots 0.8367106 22
## 23 omicron theme 0.8358337 23
## 24 omicron cov2 0.8354232 24
## 25 omicron emerge 0.8332907 25
## 26 omicron 617 0.8317066 26
## 27 omicron emergent 0.8308727 27
## 28 omicron evolved 0.8293188 28
## 29 omicron epidemics 0.8276324 29
## 30 omicron mutations 0.8270133 30“model”という単語と類似度が高い単語を30個リストアップする。
nn <- predict(model, c("model"), type = "nearest", top_n = 30)
nn## $model
## term1 term2 similarity rank
## 1 model algorithm 0.9146559 1
## 2 model stacking 0.9143786 2
## 3 model models 0.9107108 3
## 4 model oga 0.9090107 4
## 5 model configuration 0.9041815 5
## 6 model ensemble 0.9033030 6
## 7 model gpr 0.9030290 7
## 8 model classifier 0.9019660 8
## 9 model avedl 0.8963149 9
## 10 model dcnn 0.8960069 10
## 11 model gbts 0.8959891 11
## 12 model catboost 0.8930810 12
## 13 model dws 0.8903774 13
## 14 model xgboost 0.8893422 14
## 15 model tunes 0.8878289 15
## 16 model cgenet 0.8865037 16
## 17 model lightefficientnetv2 0.8841574 17
## 18 model stacked 0.8827171 18
## 19 model best 0.8810557 19
## 20 model learner 0.8806218 20
## 21 model hybrid 0.8787336 21
## 22 model cubist 0.8771569 22
## 23 model lgbm 0.8756057 23
## 24 model df 0.8753271 24
## 25 model bls 0.8746838 25
## 26 model udl 0.8738938 26
## 27 model hyperparameters 0.8729782 27
## 28 model arimax 0.8706847 28
## 29 model prediction 0.8693837 29
## 30 model serp 0.8678178 30次に、タイトルでよく用いられていた（頻出単語上位100個）について、ベクトル空間上での位置関係をt-SNE法によって視覚化する。
selector <- names(word.top100) %in% rownames(as.matrix(model))
dm <- as.matrix(model)[names(word.top100)[selector],]
word <- rownames(dm)
tsne.dm <- Rtsne(dm)視覚化
df <- data.frame(tsne.dm$Y, word)
plot_ly(df, x = ~X1, y = ~ X2, type = "scatter", mode = "text", text = word)次に、得られた単語の関係をもとに、文書を30次元の空間に埋め込む。
x <- data.frame(doc_id = titl, text = abst)
emb <- doc2vec(model, x, type = "embedding")omicronに関する説明（WHOのホームページより）をクエリにして、類似度の文書を得る。 https://www.who.int/news/item/28-11-2021-update-on-omicron
q <- txt_clean_word2vec("On 26 November 2021, WHO designated the variant B.1.1.529 a variant of concern, named Omicron, on the advice of WHO’s Technical Advisory Group on Virus Evolution (TAG-VE). This decision was based on the evidence presented to the TAG-VE that Omicron has several mutations that may have an impact on how it behaves, for example, on how easily it spreads or the severity of illness it causes. Here is a summary of what is currently known. ")
# from https://www.who.int/news/item/28-11-2021-update-on-omicron
newdoc <- doc2vec(model, q)
sim <- word2vec_similarity(emb, newdoc)
names(sim) <- rownames(emb)
sort(sim, decreasing = T)[1:20]## The COVID-19 pandemic: prediction study based on machine learning models.
## 0.9925448
## Pandemic coronavirus disease (Covid-19): World effects analysis and prediction using machine-learning techniques.
## 0.9881864
## Chemo-Preventive Effect of Vegetables and Fruits Consumption on the COVID-19 Pandemic.
## 0.9880966
## Role of Imaging and AI in the Evaluation of COVID-19 Infection: A Comprehensive Survey.
## 0.9876175
## Advanced Deep Learning Algorithms for Infectious Disease Modeling Using Clinical Data: A Case Study on COVID-19.
## 0.9872560
## A comprehensive review on variants of SARS-CoVs-2: Challenges, solutions and open issues.
## 0.9869946
## Regressive Class Modelling for Predicting Trajectories of COVID-19 Fatalities Using Statistical and Machine Learning Models.
## 0.9868742
## COVID-19 in Bangladesh: A Deeper Outlook into The Forecast with Prediction of Upcoming Per Day Cases Using Time Series.
## 0.9867825
## Real-time measurement of the uncertain epidemiological appearances of COVID-19 infections.
## 0.9867129
## Prediction and forecasting of worldwide corona virus (COVID-19) outbreak using time series and machine learning.
## 0.9866893
## Covidex: An ultrafast and accurate tool for SARS-CoV-2 subtyping.
## 0.9864149
## Coronaviruses and people with intellectual disability: an exploratory data analysis.
## 0.9863524
## Variant-driven early warning via unsupervised machine learning analysis of spike protein mutations for COVID-19.
## 0.9862817
## Diagnosis of COVID-19 and non-COVID-19 patients by classifying only a single cough sound.
## 0.9862552
## A Hybrid Protocol for Identifying Comorbidity-Based Potential Drugs for COVID-19 Using Biomedical Literature Mining, Network Analysis, and Deep Learning.
## 0.9859721
## Prediction of COVID-19 Pandemic in Bangladesh: Dual Application of Susceptible-Infective-Recovered (SIR) and Machine Learning Approach.
## 0.9857951
## Analysis on novel coronavirus (COVID-19) using machine learning methods.
## 0.9857901
## MonkeyPox2022Tweets: A Large-Scale Twitter Dataset on the 2022 Monkeypox Outbreak, Findings from Analysis of Tweets, and Open Research Questions.
## 0.9857859
## Origin of novel coronavirus causing COVID-19: A computational biology study using artificial intelligence.
## 0.9855705
## Time series forecasting of COVID-19 transmission in Canada using LSTM networks.
## 0.9855676COVID-19に関するニューラルネットワークの論文をクエリにして類似の文書を得る。 https://pubmed.ncbi.nlm.nih.gov/34745319/
my.abst <- txt_clean_word2vec("Recently, people around the world are being vulnerable to the pandemic effect of
the novel Corona Virus. It is very difficult to detect the virus infected chest
X-ray (CXR) image during early stages due to constant gene mutation of the
virus. It is also strenuous to differentiate between the usual pneumonia from
the COVID-19 positive case as both show similar symptoms. This paper proposes a
modified residual network based enhancement (ENResNet) scheme for the visual
clarification of COVID-19 pneumonia impairment from CXR images and
classification of COVID-19 under deep learning framework. Firstly, the residual
image has been generated using residual convolutional neural network through
batch normalization corresponding to each image. Secondly, a module has been
constructed through normalized map using patches and residual images as input.
The output consisting of residual images and patches of each module are fed into
the next module and this goes on for consecutive eight modules. A feature map is
generated from each module and the final enhanced CXR is produced via
up-sampling process. Further, we have designed a simple CNN model for automatic
detection of COVID-19 from CXR images in the light of 'multi-term loss' function
and 'softmax' classifier in optimal way. The proposed model exhibits better
result in the diagnosis of binary classification (COVID vs. Normal) and
multi-class classification (COVID vs. Pneumonia vs. Normal) in this study. The
suggested ENResNet achieves a classification accuracy 99.7% and 98.4% for binary
classification and multi-class detection respectively in comparison with
state-of-the-art methods.")
# Ghosh and Ghosh (2022) ENResNet: A novel residual neural network for chest X-ray enhancement based COVID-19 detection. Biomed Signal Process doi: 10.1016/j.bspc.2021.103286.
# PMID: 34745319
newdoc <- doc2vec(model, my.abst)
sim <- word2vec_similarity(emb, newdoc)
names(sim) <- rownames(emb)
sort(sim, decreasing = T)[1:20]## COVID-19 Detection Based on Image Regrouping and Resnet-SVM Using Chest X-Ray Images.
## 0.9947877
## COVID-19 deep classification network based on convolution and deconvolution local enhancement.
## 0.9944000
## An automated and fast system to identify COVID-19 from X-ray radiograph of the chest using image processing and machine learning.
## 0.9930635
## Using handpicked features in conjunction with ResNet-50 for improved detection of COVID-19 from chest X-ray images.
## 0.9930370
## CAD systems for COVID-19 diagnosis and disease stage classification by segmentation of infected regions from CT images.
## 0.9925620
## Detection and classification of lung diseases for pneumonia and Covid-19 using machine and deep learning techniques.
## 0.9924851
## Segmenting lung lesions of COVID-19 from CT images via pyramid pooling improved Unet.
## 0.9921300
## Multi-task multi-modality SVM for early COVID-19 Diagnosis using chest CT data.
## 0.9917714
## Fast and Accurate Detection of COVID-19 Along With 14 Other Chest Pathologies Using a Multi-Level Classification: Algorithm Development and Validation Study.
## 0.9917466
## Classification of COVID-19 chest X-Ray and CT images using a type of dynamic CNN modification method.
## 0.9910857
## Improving the performance of CNN to predict the likelihood of COVID-19 using chest X-ray images with preprocessing algorithms.
## 0.9909081
## Fully automatic pipeline of convolutional neural networks and capsule networks to distinguish COVID-19 from community-acquired pneumonia via CT images.
## 0.9908202
## A deep learning based approach for automatic detection of COVID-19 cases using chest X-ray images.
## 0.9908165
## [Research on coronavirus disease 2019 (COVID-19) detection method based on depthwise separable DenseNet in chest X-ray images].
## 0.9907997
## Multi-branch fusion auxiliary learning for the detection of pneumonia from chest X-ray images.
## 0.9906843
## FAM: focal attention module for lesion segmentation of COVID-19 CT images.
## 0.9906176
## FBSED based automatic diagnosis of COVID-19 using X-ray and CT images.
## 0.9905279
## An optimized KELM approach for the diagnosis of <i>COVID-19</i> from 2D-SSA reconstructed CXR Images.
## 0.9904189
## COVID-19 Detection from Chest X-ray Images Using Feature Fusion and Deep Learning.
## 0.9903693
## COVID Detection From Chest X-Ray Images Using Multi-Scale Attention.
## 0.9903296