A comprehensive bibliometric survey of micro-expression recognition system based on deep learning

Heliyon 10 (2024) e27392 A ailable online 4 Ma ch 2024 2405-8440/© 2024 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/). Re iew a icle A comp ehensi e bibliome ic su ey o mic o-exp ession ecogni ion sys em based on deep lea ning Adnan Ahmad a , Zhao Li a , ** , Shee az Iqbal b , * , Muhammad Au angzeb c , I an Ta iq a , Ayman Flah d , e , , g , Voj ech Blazek h , Lukas P okop h a Key Labo a o y o Unde wa e Acous ic Signal P ocessing o Minis y o Educa ion, School o In o ma ion Science and Enginee ing, Sou heas Uni e si y, Nanjing, 210096, China b Depa men o Elec ical Enginee ing, Uni e si y o Azad Jammu and Kashmi , Muza a abad, 13100, AJK, Pakis an c School o Elec ical Enginee ing, Sou heas Uni e si y, Nanjing, 210096, China d College o Enginee ing, Uni e si y o Business and Technology (UBT), Jeddah, 21448, Saudi A abia e MEU Resea ch Uni , Middle Eas Uni e si y, Amman, Jo dan The P i a e Highe School o Applied Sciences and Technology o Gabes, Uni e si y o Gabes, Gabes, Tunisia g Na ional Enginee ing School o Gabes, Uni e si y o Gabes, Gabes, 6029, Tunisia h ENET Cen e, VSB—Technical Uni e si y o Os a a, Os a a, Czech Republic ARTICLE INFO Keywo ds: Bibliome ic analysis Mic o exp ession Scopus Web o science ABSTRACT Mic o-exp essions (ME) a e apidly occu ing exp essions ha e eal he ue emo ions ha a human being is ying o hide, co e , o supp ess. These exp essions, which e eal a pe son’s ac ual eelings, ha e a b oad spec um o applica ions in public sa e y and clinical diagnosis. This s udy p o ides a comp ehensi e e iew o he a ea o ME ecogni ion. A bibliome ic and ne wo k analysis echniques is used o compile all he a ailable li e a u e ela ed o ME ecogni ion. A o al o 735 publica ions om he Web o Science (WOS) and Scopus da abases we e e alua ed om Decembe 2012 o Decembe 2022 using all ele an keywo ds. The i s ound o da a sc eening p oduced some basic in o ma ion, which was u he ex ac ed o ci a ion, coupling, co-au ho ship, co-occu ence, bibliog aphic, and co-ci a ion analysis. Addi ionally, a hema ic and desc ip i e analysis was execu ed o in es iga e he con en o p io esea ch indings, and esea ch echniques used in he li e a u e. The yea wise publica ions indica ed ha he published li e a u e be ween 2012 and 2017 was ela i ely low bu howe e by 2021, a nea ly 24- old inc emen made i o 154 publica ions. The h ee opmos p oduc i e jou nals and con e ences included IEEE T ansac ions on A ec i e Compu ing (n =20 publica ions) ollowed by Neu o- compu ing (n =17) and Mul imedia ools and applica ions (n =15). Zhao G was he mos p o icien au ho wi h 48 publica ions and he op in luen ial coun y was China (620 publica- ions). Publica ions by ci a ions showed ha each o he au ho s acqui ed ci a ions anging om 100 o 1225. While publica ions by o ganiza ions indica ed ha he Uni e si y o Oulu had he mos published pape s (n =51). Deep lea ning, acial exp ession ecogni ion, and emo ion ecogni ion we e among he mos equen ly used e ms. I has been disco e ed ha ME esea ch was p ima ily classi ied in he discipline o enginee ing, wi h mo e con ibu ion om China and Malaysia compa a i ely. * Co esponding au ho . ** Co esponding au ho . E-mail add esses: [email p o ec ed] (S. Iqbal), [email p o ec ed] (M. Au angzeb), [email p o ec ed] (A. Flah), [email p o ec ed] (V. Blazek), [email p o ec ed] (L. P okop). Con en s lis s a ailable a ScienceDi ec Heliyon jou nal homepage: www.cell.com/heliyon h ps://doi.o g/10.1016/j.heliyon.2024.e27392 Recei ed 12 No embe 2023; Recei ed in e ised o m 21 Feb ua y 2024; Accep ed 28 Feb ua y 2024 Heliyon 10 (2024) e27392 2 1. In oduc ion Mic o-exp essions (ME) a e ansien , low-in ensi y acial exp essions ha occu s when people y o conceal hei genuine eelings, ei he pu pose ully o subconsciously [1]. As a esul , i is di icul o disce n such genuine emo ional in o ma ion. One o he mos signi ican conce ns is ha he du a ion o a ME is exceedingly sho , las ing app oxima ely 0.04s–0.2s [2], and o he li e a u es ha e ound ha he du a ion o a ME is less han 0.33s and does no su pass 0.5s [3]. The apid a i al and disappea ance o ME makes acking and iden i ica ion mo e di icul . Simila o low-in ensi y ME, hey only apply o a po ion o acial exp ession ac ion uni s [4]. As a esul , e en wi hou o mal ins uc ion, he human eye can quickly ecognize ME and mos pa icipan s s uggle o iden i y i . Many esea che s ha e been wo king ha d o e he pas decades o assis compu e s be e in e p e human acial ME and emo ional engagemen . Takalka e al. [5], documen ed he li e a u e abou ace de ec ion and ecogni ion, il e ing, acial s uc u e de ec ion and selec ion, and classi ica ion. These me hods we e ca ego ized in o h ee ypes: appea ance-based me hods, mo ion-based me hods, and deep lea ning-based me hods. F om which i can be obse ed ha hese echniques ocused on appea ance ha e ga ne ed g ea e in e es in he li e a u e. These me hods desc ibe he dynamics o exp ession in ensi y as well as ex u al de ails like u ows and w inkles. To cha ac e ize ME in appea ance-based me hods, a a ie y o ep esen a ions ha e been used, such echniques include Gabo il e s, 2D Gabo il e and Spa se Rep esen a ion (2DGSR), Disc iminan Tenso Subspace Analysis (DTSA) and Local Bina y Pa e ns on Th ee O hogonal Planes (LBPTOP) [6–12]. Se e al s udies [11,13–16] p esen ed LBPTOP a ian s o apidly ampli y ME ecogni ion. Guo e al. [13], in oduced Th ee O hogonal Planes wi h Cen alized Bina y Pa e ns (CBP-TOP). In hei wo k hey only conside ed pixels wi h he highes weigh among hei neighbo s. Huang e al. [17], p oposed Spa io-Tempo al Local Bina y Pa e ns wi h In eg al P ojec ion. They employed in eg al p ojec ion o main ain acial image shape a ibu es and hence imp o e mic o-exp ession disc imina ion. Huang e al. [14], wen e en u he , p oposing he Spa io-Tempo al Comple ed Local Quan ized Pa e ns (STCLQP). Ins ead o homogeneous pa e ns, hey used o ien a ion, sign, and magni ude componen s o cons uc app op ia e pa e n codes. The usage o hese pa e n codes can esul in imp o ed pe o mance; howe e , i is based on he aining da ase used o de elop classi ie s. LBPTOP was used by Zong e al. [15], o de ec ME, he da a se was p ep ocessed wi h Eule ian ideo magni i- ca ion (EVM), which in ol ed ampli ying ME a low in ensi y. Liong e al. [16], s a ically selec ed di e en ace a eas based on he equency o occu ence o Ac ion Uni s egion-o -in e es (ROI-selec i e). Acco ding o Zong e al. [15], he adi ional spa ial di ision me hods canno gua an ee such accep able sub egions since i s g id size is ixed. While using he la ge sub egions may esul in noisy da a ha can in e e e wi h spa io empo al ea u e pe o mance, while using small sub egions esul s in a loss o aluable in o ma ion. They de eloped a hie a chical di ision me hod in which ace was di ided in o sub egions o a ied dimensions o add ess his issue. STLBP-IP was used o ME ea u es de ec ion [17]. ME samples was used o imp o e ecogni ion o he h ee exp essions namely "disgus ", "happiness", and "su p ise" o add ess he sca ci y o labeled ME ins ances. A Singula Value Decomposi ion (SVD) app oach was used o implemen LBP ( eps. LBP-TOP) in o de o cons uc a mac o- o-mic o ans o ma ion model o he de ec ion o ace mac o-exp ession ( eps. ME) ea u es while Coupled Me ic Lea ning (CML) me hod was used o ep esen he common cha ac e is ics o ME samples and acial mac o-in o ma ion in hei Ho Wheel Pa e ns (HWP) and HWP-TOP algo i hms o modeling ace mac o and ME [18,19]. The esul s we e hen compa ed o cu ing-edge me hodologies and i was disco e ed ha using mac o-exp ession da a o de ec ME, did no esul in a subs an ial boos in e icacy. In pas ew yea s, nume ous a icles ha e been published which discussed he mo ion-based app oaches o measu ing o ien a ions o he acial componen and non- igid mo emen [20–24]. To compu e he acial mo ion i s adjus men s among consecu i e ames he usage o he b igh ness conse a ion idea, The His og am o O ien ed Op ical Flow (HOOF) and i s a ian s we e commonly used in mo ion-based au oma ed algo i hms o ME de ec ion. To de ec mo emen changes in ME, Li e al. [24], de eloped he His og am o Image G adien O ien a ion on Th ee O hogonal Planes (HIGO-TOP). Magni ica ion on ames was used in hei p oposed app oaches o make he ea u e mo e use ul. F om las ew yea s o o e come hese p oblems, deep lea ning (DL) based me hod has been p opose which achie ed high-le el ea u e iden i ica ion and pa e n ecogni ions simul aneously, he p oposed app oaches we e assessed u ilizing a ange o deep neu al ne wo k (DNN) opologies, including he Con olu ional Neu al Ne wo k (CNN) and he Recu en Neu al Ne wo k (RNN). Fo his ins ance, Kim e al. [25], employ DL based algo i hms o ecognize ME. In hei p oposed app oach, CNN was used o iden i y spa ial ea u e space on selec ideo ames a di e en exp ession le els (onse , apex, o se ). Fu he , RNN-de i ed Long sho - e m memo y (LSTM) ne wo k was used o ecognize ime-dependen ea u es in ideo sequences. Simila ly, Peng e al. [26], came up wi h a new ype o neu al ne wo k called he "DTSCNN" which s ands o Dual Tempo al Scale Con olu ional Neu al Ne wo k o ME ecogni ion. In hei me hod o calcula ing he high-dimensional spa io empo al space o ea u es, hey inco po a ed op ical- low sequences o e a ious empo al scales in o he DTSCNN. In addi ion, Wang e al. [27], and Reddy e al. [28], p oposed an a chi- ec u e o 3D-CNNs de i ed om ideo sequences o he pu pose o con olu ional and ea u e iden i ica ion based on he 3D ke nel. Takalka e al. [29], ecen ly p esen ed a hyb id echnique o ecognizing ME based on handmade and deep ea u es. The handmade ea u es, which we e based on he LBP-TOP, depic ed he ace’s spa io- empo al mo ions, whils he deep ea u es we e de i ed using CNN. The p ocess o ecogni ion was pe o med in a “black box” using Deep Lea ning (DL) algo i hms. Al hough he de elopmen o DL based algo i hms and ou s anding classi ie s in ME iden i ica ion [30,31], hei dependabili y in ME de ec ion emains a challenge due o he limi ed numbe o ME da ase s. To add ess he issue o insu icien samples in he ME da ase , Zhi e al. [32], and Wang e al. [33], used ans e lea ning. They use he ME da ase o ain hei model. The model was hen ans e ed and ine- uned o ecognize ME. The undamen al issue wi h T ans e lea ning-based me hods is abou pe o mance which su e s as a esul o noise such as b igh ness, misaligned ace, and he ela i ely b ie du a ion and sub le mo emen o ME [34]. A. Ahmad e al. Heliyon 10 (2024) e27392 3 The eno mous size o he ea u e space is a signi ican d awback o appea ance-based app oaches. Fo ins ance, Huang e al. [14], used mo e han 23,000 ea u es while he me hod ha Wang e al. [11], de eloped, con ained abou 4425 ea u es. Wang e al. [34], used Facial Ac ion Coding Sys em (FACS) [35] o es ablish 16 ROIs in o de o sol e his issue. Addi ionally, ROIs we e also used o ME de ec ion by Liu e al. [20], and Zong e al., [15]. They demons a ed ha ME changes he appea ance o only a small numbe o local and local sub egions o he ace. S udies ha e shown ha a model’s ME e iciency can be inc eased by including ROI-based ea u es. The py amid o uni o m local bina y pa e ns was hen used as he basis o a model de eloped by Abdallah e al., [36]. I is no ewo hy ha e en a e ROIs a e used o iden i y he ea u e space o appea ance-based me hods, hei e ec i eness in di e en ia ing ME is s ill limi ed when compa ed o deep lea ning-based me hods and mo ion-based me hods. We belie e ha his is due o con usion in he classi ica ion o iden ical ME, which is one o he mos di icul challenges o all esea che s in his ield. In ligh o he p e iously discussed li e a u e, he e was a sudden o shoo in esea ch and unde s anding in he ME ecogni ion domain. As a esul , he bibliome ic analysis is conduc ed, which aids in a ull unde s anding and g asp o his opic. Bibliome ics is desc ibed as a ield o s udy in which scien is s e alua e bibliog aphic da a (i.e., published li e a u e) using a ious s a is ical and ma hema ical echniques o disco e signi ican ends and pa e ns [37–40]. Bibliome ic s udies assess and measu e he in luence and impac o publica ions. These esea che s commonly seek ou pa e ns in he p oduc ion, dissemina ion, and ecep ion o in o ma ion, alongside examining he ela i e signi icance and impac o di e se bibliog aphic elemen s such as ci a ions, au ho s, ins i u ions, and jou nals, among o he ac o s [37,38]. A comp ehensi e and unbiased e alua ion o he s a us o a speci ic ield is a undamen al objec i e o bibliome ics esea ch. This objec i e is accomplished by quan i ying he olume and calibe o ele an publica ions, as well as he equency wi h which hey a e acknowledged and u ilized by o he schola s. Such assessmen s can be u ilized by esea che s, poli- cymake s, and unding o ganiza ions o iden i y a eas o excellence and a eas in need o imp o emen , enabling hem o make well-in o med decisions ega ding he alloca ion o esou ces and suppo . Addi ionally, bibliome ics esea ch can be used o iden i y new ends and po en ial a eas o expansion in a pa icula ield by examining he shi in publishing and ci a ion p ac ices o e ime. This can help esea che s unco e new p ospec s and subjec s while also s aying up o da e on de elopmen s in hei espec i e ields. Bibliome ic s udies can also be used o examine he each and in luence o speci ic publica ions, au ho s, and ins i u ions. These esea ches ack he numbe o ci a ions, he jou nals in which he publica ions we e published, and he na ions and ins i u ions ha ci e he pape s. Such e alua ions can help esea che s e alua e hei own wo k while also unco e ing p ospec i e collabo a ions and ne wo king oppo uni ies wi h o he schola s. Fu he mo e, co po a ions can use his ype o analysis o make educa ed decisions and assess he e ec i eness o speci ic o ganiza ions, coun ies, and publica ions. Some o he p ominen so wa e p og ams used by esea che s o unde ake bibliome ic analysis include VOS iewe [41], BibExcel [42], Ci eSpace [43], and Bibliome ix [44]. Thus, p o iding da a on an open access po al o bibliome ic analysis is a solid p ac ice o suppo ing new esea che s. I is impe a i e o new esea che s o emain ab eas o de elopmen s in hei ield by p o iding pe inen analyses, e idence-based desc ip ions, and p ecise ep esen a ions de i ed om da a ob ained om he Scopus and Web o Science da abases [45]. The pu pose o his analysis is o ack he e olu ion o esea ch ends in ME o e ime. This includes he iden i ica ion o eme ging opics, shi s in ocus, changes in me hodologies. Analyzing bibliome ic da a can e eal he mos p ominen jou nals, con e ence, au ho s o ano he ME esea ch is dissemina ed. This in o ma ion can be aluable o esea che s looking o a ge speci ic opic in ME ecogni ions. This s udy’s indings can be u ilized o iden i y cu en end and ho opics o published a icles o u he s udy and esea ch. This also p o ide in-dep h unde s anding o a esea ch opic by inducing and mapping au ho s’ egions. To gain esea ch knowledge, i is necessa y o be awa e o he exis ing cons ain s [46]. This helps disco e ing new ends as well as exposing he u u e esea ch, which is he undamen al mo i a ion o his s udy. To he bes o ou knowledge, no p e ious wo k e alua ed he ex ensi e bibliome ic li e a u e e alua ion and so wa e-based hema ic analysis on mic o-exp ession ecogni ion. Fu he mo e, his s udy p esen ed a amewo k o hema ic analysis ha mos esea che s and indus ial expe s can employ in u u e esea ch. The goal o his s udy was o p o ide answe s o he ollowing ques ions. •Wha a e he esea ch ho spo s o ME ecogni ion in he exis ing li e a u e? •Which coun ies, ME- ela ed a icles, au ho s, and/o jou nals ha e achie ed ema kable esul s? •Wha a e he key opics and ad ancemen s in he ME ecogni ion esea ch? •Wha a e he limi a ions and esea ch p io i ies in conside a ion o ME ecogni ion? In o de o ca y ou an in-dep h bibliome ics li e a u e e iew, R-based ool was used in his s udy as well as he VOSViewe so wa e. As a esul , his s udy was a emp ed o imp o e he ollowing a eas o he li e a u e now in exis ence. •Ou wo k has es ablished a hema ic amewo k h ough he u iliza ion o so wa e-based hema ic analysis, which has acili a ed he in es iga ion o main esea ch ho spo . •In his s udy, we addi ionally classi ied he p eceding esea ch in es iga ions based on he employed me hodology. The esul p o ided an insigh in o cu en esea ch app oaches as well as app oaches ha ha e been o e looked. The a icle is s uc u ed as ollows, Sec ion 2 ou lines he me hodology, da a sou ces, and ools used in he bibliome ics analysis esea ch. Sec ion 3 p esen s he esul s and obse a ions, which p o ides a comple e bibliome ic s udy in e ms o pe o mance analysis and science mapping. Sec ion 4 p esen s some limi a ions o he s udy. Sec ion 5 p esen s he conclusions o he esea ch and summa izes i s indings. A. Ahmad e al. Heliyon 10 (2024) e27392 4 2. Ma e ials and me hods 2.1. Resea ch me hodology The goal o his s udy was o pe o m a bibliome ic analysis abou ME ecogni ion. The p oposed s udy e alua ed published a icles by ex ac ing quan i a i e in o ma ion om g aphical da a using a ange o s a is ical echniques (VOS iewe , an R-based ool). Da a was ga he ed using a ious sea ch engines and analyzed o es ima e he numbe o published a icles each yea as well as ele an s udy subjec s. I also included in o ma ion abou p ominen au ho s, coun ies, o ganiza ions, and publica ions in he ield o ME ecogni ion. Fu he mo e, depending on he s a is ics, he mos in luen ial subjec a eas and a icles we e highligh ed. As a esul , ou planned bibliome ic s udy has included wo ks on mic o acial exp ession and ecogni ion since he commencemen o his su ey. 2.1.1. Da a collec ion In o de o ob ain bibliome ic da a, sea ch e ms we e used in he WOS and Scopus da abases on Jan 10, 2023, and he e ie ed esul s we e u he anked based on he i le, abs ac , and o he inclusion/exclusion c i e ia sugges ed by a ious in es iga o s. 2.1.2. Iden i ica ion o sea ch e ms To disco e he ele an e ms, se e al publica ions om he p e ious li e a u e we e e iewed. To loca e ele an ph ases, Google Schola was also sea ched o "mic o-exp ession" and " acial mic o-exp ession." The keywo ds " acial mic oexp ession" OR "mic o- exp ession" OR "mic o exp ession" OR " acial-mic oexp ession ". A p elimina y sea ch was hen conduc ed using he i le, abs ac and keywo d o he publica ion. This esul ed in he de ec ion o 782 a icles in Scopus da abase while he same keywo ds we e also used in WOS and a o al o 428 publica ions we e e ie ed. The sea ch pe iod was hen limi ed o 2012-22 wi h only included English- language and pee - e iewed open access wo ks and he inal esul s ob ained om Scopus we e 712 publica ions while ha o WoS we e 404 publica ions. The sea ch e ms we e hen employed o acqui e esea ch and e iew a icles om bo h WOS and Scopus da abases as o Janua y 10 h, 2023. 2.1.3. A icle selec ion The selec ion o an app op ia e da abase is c i ical in conduc ing e icien sea ches. We made conside able use o he Scopus and WOS da a bases in his espec . Scopus, in pa icula , gi es academics g ea e isibili y han o he da abases, and i s sophis ica ed sea ch unc ionali y ools allow hem o o ganize e e ences and collec ci a ions om pape s. Fu he mo e, he Scopus da abase includes jou nals om a wide ange o publishe s, including Sp inge , Else ie , Taylo & F ancis, Eme ald Insigh , and IEEE [43,47]. In addi ion, WOS was used o ill in he missing a icles and jou nals in Scopus. The i s s ep was o sea ch bo h da abases o ele an Fig. 1. A icle selec ion p ocedu e used in his s udy. A. Ahmad e al. Heliyon 10 (2024) e27392 5 ph ases. Abs ac s, i les, and keywo ds yielded 712 o al esul s o Scopus and 404 o WOS. Fig. 1 depic ed he a icle segmen p ocess employed in his in es iga ion. 712 Scopus a icles and 404 WOS a icles we e ex ac ed in wo di e en o ma s. A e combining he wo da a iles, R-s udio was used o loca e and emo e iden ical a icles. As a esul , 735 a icles in Comma Sepa a ed Values (CSV) o ma we e expo ed o bibliome ic analysis. 3. Resul s and discussion 3.1. Bibliome ic analysis This sec ion add esses he bibliome ic analysis o 735 selec ed publica ions’ au ho keywo ds, au ho pa ne ships, jou nals, ci- a ions, ins i u ions, heme e alua ion, and bibliog aphic coupling. 3.1.1. Publica ion by yea Fig. 2 depic ed a signi ican inc ease in he publica ion o esea ch a icles o e he las en yea s, indica ing inc eased in e es in mic o-exp ession ecogni ion in academic communi ies. Fig. 2 showed ha he published li e a u e was ela i ely low be ween 2012 and 2017. In 2012, he e we e only 12 published pape s; by 2021, a nea ly 24- old inc ease made i o 154 publica ions. We an icipa ed ha he publica ion o mic o-exp ession ecogni ion will g ow exponen ially in 2023 and subsequen yea s. 3.1.2. Publica ion by jou nals, au ho s and coun ies Fig. 3 showed he op i e jou nals, i is impe a i e o conside he cumula i e quan i y o esea ch pape s ha ha e been published wi hin he pas decade. The h ee mos p oduc i e jou nals and con e ences, as shown in Fig. 3a, we e IEEE T ansac ions on A ec i e Compu ing (20 pape s), Neu ocompu ing (17 pape s), Mul imedia ools and applica ions (15 pape s). I was obse ed ha he op h ee jou nals oge he published 41% o he o e all li e a u e. Fig. 3b illus a ed he op ou au ho s wi h he mos a icles published be ween 2012 and 2022. In addi ion, he igu e e ealed ha Zhao G was he mos p o icien au ho (48 publica ions) ollowed by Wang S and NA N wi h 47 and 37 publica ions. Likewise, he op in luen ial coun y as shown in Fig. 3c was China (620 publica ions) ollowed by Malaysia (91), and Finland (60) in e ms o scien i ic p oduc ion. 3.1.3. Publica ions by ci a ions The p esen s udy e alua ed he agg ega e numbe o ci a ions while ga he ing da a and insigh s on no ewo hy au ho s in he domain o mic o-exp ession ecogni ion. The op en mos equen ly ci ed a icles om he Scopus and WOS da abases we e sum- ma ized in Fig. 4. Fig. 4 illus a ed ha each o he au ho s acqui ed ci a ions anging om 100 o 1225. I is impo an o no e ha due o a ia ions in indexing algo i hms and ime pe iods employed by di e en da abases, he o al numbe o ci a ions ob ained om Google Schola and o he da abases may di e . Acco ding o Fig. 4, he au ho Zhao Guoyang’s wo ks ob ained he mos ci a ions (1225). She is cu en ly a ull P o esso a he Uni e si y o Oulu’s Cen e o Machine Vision and Signal Analysis (IAPR Fellow, IEEE Senio Membe ). The documen s o au ho Fu Xiaolan come nex , wi h a o al o 1030 ci a ions. She is a P o esso o Psychology and he Fig. 2. Publica ion o mic o exp ession ecogni ion pe yea . A. Ahmad e al. Heliyon 10 (2024) e27392 6 Di ec o o he Chinese Academy o Sciences’ Ins i u e o Psychology. She is a pionee in China’s mic o-exp ession s udy. She has c ea ed h ee open-access mic o-exp ession da abases: CASME, CASME II, and CAS (MEU2). He esea ch in e es s include lie de ec ion, emo ional compu ing, lea ning, pe cep ion and a en ion, and pe cep ion and a en ion. Ano he au ho wi h a high numbe o ci a ions is Wang S. His wo k has been ci ed 1006 imes, and he is now an Assis an Resea che a he Ins i u e o Psychology o he Chinese Academy o Sciences. He is he au ho o app oxima ely 40 schola ly pape s. A he 2011 In e na ional Join Con e ence on Biome ics, he is among he en doc o al conso ium pa icipan s. The e o e, i is accep able o belie e ha all he publica ions depic ed in Fig. 4 we e among he mos well-known s udies ega ding ecogni ion o mic o-exp essions in li e a u e. 3.1.4. Publica ion by ins i u ions Fig. 5 depic ed ha he Uni e si y o Oulu has he mos published pape s (51), among he op nine he op nine academic in- s i u ions. The Ins i u e o Psychology ( he Chinese Academy o Sciences) was e ealed o be he second anked wi h o al o 47 publica ions while he Mul imedia Uni e si y (Malaysia) was anked hi d wi h 39 numbe o publica ions. Acco ding o he analyzed esul s, se en o he op en mos p oduc i e ins i u ions we e om China and ac i ely pu sued esea ch in o mic o-exp ession ecogni ion. 3.1.5. Common wo ds in i le and abs ac Fig. 6 illus a ed he Common wo ds used in abs ac s and i les o all he publica ions. The mos common e m used was mic o- exp essions wi h 208 imes in i le-abs ac s, ollowed by acial exp essions (84 imes), ace ecogni ion (74 imes), and deep lea ning (73 imes). 3.1.6. Common wo ds in keywo ds The wo ds mos commonly employed by esea che s in he keywo ds sec ion we e also assessed, gi en hei endency o con ey he cen al ocus o he esea ch. Fig. 7a depic ed ha deep lea ning, acial exp ession ecogni ion, and emo ion ecogni ion we e among he mos equen ly used e ms. The subs an ial and conspicuous on isibili y o he keywo ds is appa en , as demons a ed by hei la ge and bold on . Fig. 7b illus a ed he co esponding analysis and clus e ing dend og am o he a o emen ioned keywo ds. A dend og am is a diag am ha depic ed he wo d’s hie a chical ela ionship. A dend og am is p ima ily used o de e mine he bes way Fig. 3. (A) Publica ion by Jou nals, (B) Mos in luen ial au ho based on o al publica ions, (C) p oduc i e coun ies. Fig. 4. Publica ion by ci a ions. A. Ahmad e al. Heliyon 10 (2024) e27392 7 o assign wo ds o clus e s. Fig. 7b depic ed he hie a chical clus e ing o wo d obse a ions using a dend og am. Focusing on he heigh a which any wo wo ds we e connec ed oge he was essen ial o dend og am comp ehension. Fig. 7b showed ha mic o exp ession and acial mic o exp ession ecogni ion we e mo e simila in he b own clus e , wi h he sho es link connec ing hem. The heigh o he blue and g een clus e s, on he o he hand, was g ea e , indica ing he di e ence be ween he wo ds. 3.1.7. Ne wo k analysis This bibliome ics s udy was conduc ed using Bibliome ix (R-based so wa e o mapping li e a u e), o quan i a i e bibliome ic and scien ome ic esea ch, he bibliome ix R-package (h p://www.bibliome ix.o g) o e s a numbe o ools. I is w i en in R, an open-sou ce language, en i onmen , and ecosys em. The bes easons o choose R o e o he languages o scien i ic compu ing may be ound in i s ex ensi e, powe ul s a is ical algo i hms, easy access o excellen nume ical ou ines, and in eg a ed da a isualiza ion ools. Simila ly, VOSViewe [41] is a isualiza ion ool speci ically designed o cons uc ing and isualizing bibliome ic ne wo ks. I is widely used in he academic and esea ch communi y o analyzing ela ionships be ween schola ly en i ies such as au ho s, key- wo ds, and publica ions. The mains eam o he da a analysis was conduc ed wi h he help o Bibliome ics, based on Bibliog aphic Cooccu ence Analysis (BCA), Bibliog aphic Coupling, Co-Au ho ship, Quo a ion, and Co-Ci a ion Analysis. To analyze he da a and epo he esul s, he so wa e manuals o Gule ia and Kau [48] we e used. In ne wo k isualiza ions o scien i ic da a, a ious en i ies such as au ho s, keywo ds, na ions, and o ganiza ions ha e been depic ed by emphasizing p ominen nodes [49]. I is a a e occu ence o obse e wo en i ies, such as au ho s and publishing, being po ayed in a single map. The size o he node (ci cle) in ne wo k mapping ep esen s he measu ed alue o hese elemen s. The signi icance o he node is di ec ly p opo ional o i s nume ical alue, which p ima ily indica es he numbe o ci a ions o occu - ences in an a icle. The linkages (edges) ha connec he nodes e lec hei ela ionship. The s eng h (weigh ) alue o connec ions de e mines he associa ion be ween wo nodes, wi h a highe o al link s eng h (TLS) alue indica ing a s onge associa ion. Consequen ly, i a node ep esen s he equency o appea ance o an a icle, he links be ween nodes ep esen he numbe o e - e ences exchanged. The e o e, as he numbe o nodes inc eases, he TLS alue (weigh ) also inc eases. Addi ionally, he colo and posi ion o nodes in a ne wo k map se e as ele an indica o s. When wo a icles (nodes) a e in close p oximi y, hey appea linked and sha e a g ea e numbe o e e ences. The usage o he same colo signi ies ha he a icles belong o he same ca ego y. 3.1.8. Co-occu ence analysis (documen s) The keywo d sec ions o he au ho , o example, p o ide in o ma ion in a dis inc i e manne . The e m "co-occu ence" e e s o how equen ly a e m appea ed in pa icula publica ions [50]. The s eng h o a wo d’s o e all leng h de e mines how many imes i appea ed in a ex . The size o he node is di ec ly p opo ional o he equency o he wo ds; o ins ance, a la ge node size indica es a highe p e alence o he e m. A hicke line linking wo o mo e e ms indica es hei p oximi y o a pa icula clus e . To be e g asp he in ellec ual e ms ci ed by di e se p o esso s, co-occu ence analysis was used. In o al, 1410 au ho keywo ds and 1989 index keywo ds we e disco e ed. Table 1 displays he TLS alue o he en op au ho s and index keywo ds. "Mic o-exp ession ecogni ion, mic o-exp ession," and "deep lea ning" we e iden i ied as he mos essen ial e ms in bo h g oups. Fig. 8a depic ed he ne wo k isualiza ion o all keywo ds. Mic o-exp ession ecogni ion, mic o-exp essions, and deep lea ning we e he mos p ominen nodes. Keywo ds we e also in es iga ed be ween he yea s o 2012 and Oc obe 2022. The o e lay isualiza ion o all keywo ds is shown in Fig. 8b. A close examina ion o he g aph e ealed ha he keywo ds used mos equen ly in la e 2022 we e exp ession ecogni ion, con olu ional neu al ne wo k, mic o-exp ession analysis, ask analysis, and ea u e ex ac ions. Fig. 5. Top nine ins i u ions numbe s o publica ion. A. Ahmad e al. Heliyon 10 (2024) e27392 8 Fig. 6. T ee-map o he mos commonly used wo ds in ME ecogni ion publica ions’ i les and abs ac s. A. Ahmad e al. Heliyon 10 (2024) e27392 9 Fig. 7. (a). Wo d cloud o he mos ele an keywo ds in mic o exp ession ecogni ion publica ions (b). Dend og am o Au ho s keywo ds. Table 1 The ou comes de i ed om he co-occu ence analysis conduc ed on au ho and index keywo ds. Au ho Keywo d TLS Index Keywo d TLS Mic o-exp ession ecogni ion 203 Mic o-exp ession 1128 Deep lea ning 174 Face ecogni ion 445 Mic o-exp ession 169 Facial exp essions 434 Op ical low 125 Deep lea ning 381 Emo ion ecogni ion 84 CNN 282 Fea u es Ex ac ion 81 Compu e ision 203 Recogni ion 65 Classi ica ion 178 A ec i e compu ing 33 Lea ning sys em 181 T ans e lea ning 49 Image enhancemen 55 Task analysis 46 seman ics 47 A. Ahmad e al. Heliyon 10 (2024) e27392 16 Fig. 12. Re e ences based Co-ci a ion analysis. A. Ahmad e al. Heliyon 10 (2024) e27392 17 a ailable in da abases. Addi ionally, he use o ci a ion ne wo ks as a gauge o e ec and quali y may be subjec ed o c i icism. Th ough he comp ehensi e and sys ema ic inclusion o all documen kinds and languages, he compa a i e u iliza ion o a ious da abases, and he use o bibliome ic mapping ools, hese cons ain s could se e as u he esea ch ields o bibliome ic analysis on simila s udies. 5. Conclusion This esea ch endea o cons i u ed a comp ehensi e examina ion o published a icles pe aining o he domain o mic o exp ession ecogni ion. I encompassed undamen al da a de i ed om an eceden li e a u e, encompassing he quan i y o a icles, yea ly publica ions, and opical domains, and sc u inized he mos impac ul au ho s, a icles, jou nals, and coun ies h ough he u iliza ion o VOS iewe and R-based ools. The p oposed s udy e ealed ha he majo i y o a icles we e abou mic o exp ession and acial mic o-exp ession ecogni ion. In e ms o he subjec domain, he esea ch ho spo s we e mic o-exp ession ecogni ion and deep lea ning, ollowed by a co e applica ion o mic o-exp ession ecogni ion (op ical low, emo ion ecogni ion, ea u es ex ac ions, a ec i e compu ing, ans e lea ning). Fu he mo e, he au ho ’s keywo ds included he undamen al e ms o ME ecogni ion, whe eas he index keywo ds included b oad e ms. When i comes o jou nals, IEEE T ansac ions on A ec i e Compu ing has pub- lished he mos a icles as compa ed o he o he jou nals. In e ms o coun ies, China published he mos a icles and was also leading in e ms o au ho ship and co-ci a ion analysis using TLS and ci a ion alues. I was also no iced ha majo i y o he au ho s and ins i u ions we e om China and Malaysia, wi h mic o exp ession ecogni ion being he leading esea ch a ea. In u u e wo k, an algo i hm which ake less aining ime when ain on mic o exp ession da ase s can be de eloped. Secondly, he mic o exp ession da ase is highly imbalanced so o e coming his p oblem an algo i hm can be imp o ed o o e come his issue. CRediT au ho ship con ibu ion s a emen Adnan Ahmad: W i ing – o iginal d a , So wa e, Concep ualiza ion. Zhao Li: In es iga ion, Fo mal analysis. Shee az Iqbal: W i ing – e iew & edi ing, P ojec adminis a ion, In es iga ion. Muhammad Au angzeb: Fo mal analysis, Da a cu a ion. I an Ta iq: Visualiza ion, Valida ion. Ayman Flah: In es iga ion, Funding acquisi ion, Da a cu a ion. Voj ech Blazek: Visualiza ion, Valida ion. Lucas P okop: In es iga ion, Fo mal analysis. Decla a ion o compe ing in e es The au ho s decla e he ollowing inancial in e es s/pe sonal ela ionships which may be conside ed as po en ial compe ing in- e es s:This wo k Aknowldgmen : This pape was suppo ed by he ollowing p ojec TN02000025 Na ional Cen e o Ene gy II. The au ho s decla e he e is no con lic o in e es . Acknowledgmen This a icle has been p oduced wi h he inancial suppo o he Eu opean Union unde he REFRESH – Resea ch Excellence Fo Region Sus ainabili y and High- ech Indus ies p ojec numbe CZ.10.03.01/00/22_003/0000048 ia he Ope a ional P og amme Jus T ansi ion and pape was suppo ed by he ollowing p ojec TN02000025 Na ional Cen e o Ene gy II. Re e ences [1] W. Zheng, S. Lu, Z. Cai, R. Wang, L. Wang, L. Yin, PAL-BERT: an imp o ed ques ion answe ing model, Compu . Model. Eng. Sci. (2023), h ps://doi.o g/ 10.32604/cmes.2023.046692. [2] X. Liu, G. Zhou, M. Kong, Z. Yin, X. Li, L. Yin, W. Zheng, De eloping mul i- labelled co pus o wi e sho ex s: a semi-au oma ic me hod, Sys ems 11 (8) (2023) 390, h ps://doi.o g/10.3390/sys ems11080390. [3] W.-J. Yan, Q. Wu, J. Liang, Y.-H. Chen, X. Fu, How as a e he leaked acial exp essions: he du a ion o mic o-exp essions, J. Non e bal Beha . 37 (2013) 217–230. [4] S. Po e , L. Ten B inke, Reading be ween he lies: iden i ying concealed and alsi ied emo ions in uni e sal acial exp essions, Psychol. Sci. 19 (5) (2008) 508–514. [5] M. Takalka , M. Xu, Q. Wu, Z. Chaczko, A su ey: acial mic o-exp ession ecogni ion, Mul imed. Tool. Appl. 77 (15) (2018) 19301–19325. [6] P. Zhang, X. Ben, R. Yan, C. Wu, C. Guo, Mic o-exp ession ecogni ion sys em, Op ik 127 (3) (2016) 1395–1400. [7] S.-J. Wang, H.-L. Chen, W.-J. Yan, Y.-H. Chen, X. Fu, Face ecogni ion and mic o-exp ession ecogni ion based on disc iminan enso subspace analysis plus ex eme lea ning machine, Neu al P ocess. Le . 39 (2014) 25–43. [8] S.-J. Wang, W.-J. Yan, X. Li, G. Zhao, X. Fu, Mic o-exp ession ecogni ion using dynamic ex u es on enso independen colo space, in: 2014 22nd In e na ional Con e ence on Pa e n Recogni ion, IEEE, 2014, pp. 4678–4683. [9] W.-J. Yan, e al., Casme II: an imp o ed spon aneous mic o-exp ession da abase and he baseline e alua ion, PLoS One 9 (1) (2014) e86041. [10] Y. Wang, J. See, R.C.-W. Phan, Y.-H. Oh, Lbp wi h six in e sec ion poin s: educing edundan in o ma ion in lbp- op o mic o-exp ession ecogni ion, in: Compu e Vision–ACCV 2014: 12 h Asian Con e ence on Compu e Vision ol. 12, Sp inge , Singapo e, Singapo e, 2015, pp. 525–537. No embe 1-5, 2014, Re ised Selec ed Pape s, Pa I. [11] Y. Wang, e al., E ec i e ecogni ion o acial mic o-exp essions wi h ideo mo ion magni ica ion, Mul imed. Tool. Appl. 76 (2017) 21665–21690. [12] F. Qu, S.-J. Wang, W.-J. Yan, H. Li, S. Wu, X. Fu, Cas (ME) $^ 2$: a da abase o spon aneous mac o-exp ession and mic o-exp ession spo ing and ecogni ion, IEEE T ansac ions on A ec i e Compu ing 9 (4) (2017) 424–436. [13] Y. Guo, C. Xue, Y. Wang, M. Yu, Mic o-exp ession ecogni ion based on CBP-TOP ea u e wi h ELM, Op ik 126 (23) (2015) 4446–4451. [14] X. Huang, G. Zhao, X. Hong, W. Zheng, M. Pie ik¨ ainen, Spon aneous acial mic o-exp ession analysis using spa io empo al comple ed local quan ized pa e ns, Neu ocompu ing 175 (2016) 564–578. A. Ahmad e al. Heliyon 10 (2024) e27392 18 [15] Y. Zong, X. Huang, W. Zheng, Z. Cui, G. Zhao, Lea ning om hie a chical spa io empo al desc ip o s o mic o-exp ession ecogni ion, IEEE T ans. Mul imed. 20 (11) (2018) 3160–3172. [16] S.-T. Liong, J. See, R.C.-W. Phan, K. Wong, S.-W. Tan, Hyb id acial egions ex ac ion o mic o-exp ession ecogni ion sys em, Jou nal o Signal P ocessing Sys ems 90 (2018) 601–617. [17] X. Liu, S. Wang, S. Lu, Z. Yin, X. Li, L. Yin, W. Zheng, Adap ing ea u e selec ion algo i hms o he classi ica ion o Chinese ex s, Sys ems 11 (9) (2023) 483, h ps://doi.o g/10.3390/sys ems11090483. [18] X. Jia, X. Ben, H. Yuan, K. Kpalma, W. Meng, Mac o- o-mic o ans o ma ion model o mic o-exp ession ecogni ion, Jou nal o compu a ional science 25 (2018) 289–297. [19] X. Ben, X. Jia, R. Yan, X. Zhang, W. Meng, Lea ning e ec i e bina y desc ip o s o mic o-exp ession ecogni ion ans e ed by mac o-in o ma ion, Pa e n Recogn. Le . 107 (2018) 50–58. [20] Y.-J. Liu, J.-K. Zhang, W.-J. Yan, S.-J. Wang, G. Zhao, X. Fu, A main di ec ional mean op ical low ea u e o spon aneous mic o-exp ession ecogni ion, IEEE T ansac ions on A ec i e Compu ing 7 (4) (2015) 299–310. [21] F. Xu, J. Zhang, J.Z. Wang, Mic oexp ession iden i ica ion and ca ego iza ion using a acial dynamics map, IEEE T ansac ions on A ec i e Compu ing 8 (2) (2017) 254–267. [22] H. Lu, K. Kpalma, J. Ronsin, Mo ion desc ip o s o mic o-exp ession ecogni ion, Signal P ocess. Image Commun. 67 (2018) 108–117. [23] S.-T. Liong, J. See, K. Wong, R.C.-W. Phan, Less is mo e: mic o-exp ession ecogni ion om ideo using apex ame, Signal P ocess. Image Commun. 62 (2018) 82–92. [24] X. Li, e al., Towa ds eading hidden emo ions: a compa a i e s udy o spon aneous mic o-exp ession spo ing and ecogni ion me hods, IEEE ansac ions on a ec i e compu ing 9 (4) (2017) 563–577. [25] D.H. Kim, W.J. Badda , Y.M. Ro, Mic o-exp ession ecogni ion wi h exp ession-s a e cons ained spa io- empo al ea u e ep esen a ions, in: P oceedings o he 24 h ACM In e na ional Con e ence on Mul imedia, 2016, pp. 382–386. [26] M. Peng, C. Wang, T. Chen, G. Liu, X. Fu, Dual empo al scale con olu ional neu al ne wo k o mic o-exp ession ecogni ion, F on . Psychol. 8 (2017) 1745. [27] S.-J. Wang, e al., Mic o-exp ession ecogni ion wi h small sample size by ans e ing long- e m con olu ional neu al ne wo k, Neu ocompu ing 312 (2018) 251–262. [28] S.P.T. Reddy, S.T. Ka i, S.R. Dubey, S. Mukhe jee, Spon aneous acial mic o-exp ession ecogni ion using 3D spa io empo al con olu ional neu al ne wo ks, in: 2019 In e na ional Join Con e ence on Neu al Ne wo ks (IJCNN), IEEE, 2019, pp. 1–8. [29] M.A. Takalka , M. Xu, Z. Chaczko, Mani old ea u e in eg a ion o mic o-exp ession ecogni ion, Mul imed. Sys . 26 (2020) 535–551. [30] Y. Huang, Y. Yan, S. Chen, H. Wang, Exp ession- a ge ed ea u e lea ning o e ec i e acial exp ession ecogni ion, J. Vis. Commun. Image Rep esen . 55 (2018) 677–687. [31] Z. Yu, G. Liu, Q. Liu, J. Deng, Spa io- empo al con olu ional ea u es wi h nes ed LSTM o acial exp ession ecogni ion, Neu ocompu ing 317 (2018) 50–57. [32] R. Zhi, H. Xu, M. Wan, T. Li, Combining 3D con olu ional neu al ne wo ks wi h ans e lea ning by supe ised p e- aining o acial mic o-exp ession ecogni ion, IEICE T ans. In o Sys . 102 (5) (2019) 1054–1064. [33] C. Wang, M. Peng, T. Bi, T. Chen, Mic o-a en ion o mic o-exp ession ecogni ion, Neu ocompu ing 410 (2020) 354–362. [34] S.-J. Wang, e al., Mic o-exp ession ecogni ion using colo spaces, IEEE T ans. Image P ocess. 24 (12) (2015) 6034–6047. [35] P. Ekman, W.V. F iesen, Facial Ac ion Coding Sys em, En i onmen al Psychology & Non e bal Beha io , 1978. [36] T. Ben Abdallah, R. Gue mazi, M. Hammami, Towa ds mic o-exp ession ecogni ion h ough py amid o uni o m empo al local bina y pa e n ea u es, in: In e na ional Con e ence on In elligen Sys ems Design and Applica ions, Sp inge , 2018, pp. 629–640. [37] N. Don hu, S. Kuma , D. Mukhe jee, N. Pandey, W.M. Lim, How o conduc a bibliome ic analysis: an o e iew and guidelines, J. Bus. Res. 133 (2021) 285–296. [38] N. Don hu, W. Reina z, S. Kuma , D. Pa naik, A e ospec i e e iew o he i s 35 yea s o he in e na ional jou nal o esea ch in ma ke ing, In . J. Res. Ma ke . 38 (1) (2021) 232–269. [39] I. Dan ila-del-Valle, C. Es ´ e ez-Mendoza, F.J. La a, Human esou ces aining: a bibliome ic analysis, J. Bus. Res. 101 (2019) 627–636. [40] B. Moyle, C.-l. Moyle, L. Ruhanen, D. Wea e , A. Hadinejad, A e we eally p og essing sus ainable ou ism esea ch? A bibliome ic analysis, J. Sus ain. Tou ism 29 (1) (2020) 106–122. [41] N.J. Van Eck, L. Wal man, VOS: a new me hod o isualizing simila i ies be ween objec s, in: Ad ances in Da a Analysis: P oceedings o he 30 h Annual Con e ence o he Gesellscha ü Klassi ika ion eV, F eie Uni e si ¨ a Be lin, Sp inge , 2007, pp. 299–306. Ma ch 8–10, 2006. [42] O. Pe sson, R. Danell, J.W. Schneide , How o use Bibexcel o a ious ypes o bibliome ic analysis, Celeb a ing schola ly communica ion s udies: A Fes sch i o Olle Pe sson a his 60 h Bi hday 5 (2009) 9–24. [43] N. Ma lahi, M. Thelwall, When a e eade ship coun s as use ul as ci a ion coun s? S copus e sus M endeley o LIS jou nals, Jou nal o he Associa ion o in o ma ion Science and Technology 67 (1) (2016) 191–199. [44] C. Chen, Ci eSpace II: de ec ing and isualizing eme ging ends and ansien pa e ns in scien i ic li e a u e, J. Am. Soc. In . Sci. Technol. 57 (3) (2006) 359–377. [45] P. Tupe-Waghma e, R.R. Joshi, Bibliome ic Su ey on Mul ipu pose Face Recogni ion Sys em Using Deep Lea ning, ol. 2021, Lib a y Philosophy and P ac ice, 2021. [46] M.-L. Tseng, T.P.T. T an, H.M. Ha, T.-D. Bui, M.K. Lim, Sus ainable indus ial and ope a ion enginee ing ends and challenges owa d Indus y 4.0: a da a d i en analysis, Jou nal o Indus ial and P oduc ion Enginee ing 38 (8) (2021) 581–598. [47] Z. Liu, C. Wen, Z. Su, S. Liu, J. Sun, W. Kong, Z. Yang, Emo ion-seman ic-awa e dual con as i e lea ning o epis emic emo ion iden i ica ion o lea ne - gene a ed e iews in MOOCs, IEEE T ansac . Neu al Ne wo ks Lea n. Sys . (2023), h ps://doi.o g/10.1109/TNNLS.2023.3294636. [48] D. Gule ia, G. Kau , Bibliome ic analysis o ecop eneu ship using VOS iewe and RS udio Bibliome ix, 1989–2019, Lib . Hi Technol. 39 (4) (2021) 1001–1024. [49] M. Akinlolu, T.C. Haup , D.J. Edwa ds, F. Simpeh, A bibliome ic e iew o he s a us and eme ging esea ch ends in cons uc ion sa e y managemen echnologies, In e na ional Jou nal o Cons uc ion Managemen 22 (14) (2022) 2699–2711. [50] M. Callon, J.-P. Cou ial, W.A. Tu ne , S. Bauin, F om ansla ions o p oblema ic ne wo ks: an in oduc ion o co-wo d analysis, Soc. Sci. In . 22 (2) (1983) 191–235. [51] F.J. Acedo, C. Ba oso, C. Casanue a, J.L. Gal´ an, Co-au ho ship in managemen and o ganiza ional s udies: an empi ical and ne wo k analysis, J. Manag. S ud. 43 (5) (2006) 957–983. [52] L. Cisne os, M. Ibanescu, C. Keen, O. Loba o-Calle os, J. Niebla-Za a ain, Bibliome ic s udy o amily business succession be ween 1939 and 2017: mapping and analyzing au ho s’ ne wo ks, Scien ome ics 117 (2018) 919–951. [53] D. C ane, Social s uc u e in a g oup o scien is s: a es o he “in isible college” hypo hesis, in: Social Ne wo ks, Else ie , 1977, pp. 161–178. [54] W. Me ghani, A. Da ison, M. Yap, Facial mic o-exp essions g and challenge 2018: e alua ing spa io- empo al ea u es o classi ica ion o objec i e classes, in: 2018 13 h IEEE In e na ional Con e ence on Au oma ic Face & Ges u e Recogni ion (FG 2018), IEEE, 2018, pp. 662–666. [55] M.E. Kemeny, e al., Con empla i e/emo ion aining educes nega i e emo ional beha io and p omo es p osocial esponses, Emo ion 12 (2) (2012) 338. [56] I. Zupic, T. ˇ Ca e , Bibliome ic me hods in managemen and o ganiza ion, O gan. Res. Me hods 18 (3) (2015) 429–472. [57] B.C.A. Pe oni, E.M. de Mo aes, R.F. Gonçal es, Big da a analy ics o logis ics and dis ibu ions using blockchain, in: IFIP In e na ional Con e ence on Ad ances in P oduc ion Managemen Sys ems, Sp inge , 2018, pp. 363–369. A. Ahmad e al.