检索结果-南通市图书馆

a.toma.ic Finding of Bra.n-Tumour Group Using CNN Segmenta.ion a.d Moth-Fla.e-a.gorithm,Selected Deep a.d Ha.dcra.ted Fea.ures

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Computers, Ma.eria.s & Continua 2024年第5期79卷 2585-2608页

作者： Ima. Sa.d a. Na.mi syed a.w.e a.junid syed Junid Muha.ma. lmra. a.ma. K.Suresh Ma.ic Fa.ulty of Electronic Engineering Technology University Malaysia Perils(UniMAP)Pauh Putra CampusArauPerlis02600Malaysia Depa.tment of Electrica. a.d Communica.ion Engineering National University of Science and TechnologyAL HailAL Seeb130Sultanate of Oman Depa.tment of Resea.ch a.d Innova.ion Saveetha School of EngineeringSaveetha Institute of Medical and Technical Sciences(SIMATS)ChennaiTamiNadu602105India

a.gmenta.ion of a.norma. cells in the bra.n ca.ses bra.n tumor(BT),a.d ea.ly screening a.d trea.mentw.ll reduce its ha.shness in ***’s clinica. level screening is usua.ly performed w.th Ma.netic Resona.ce Ima.ing(MRI)due to its multi-moda.ity *** overa.l a.ms of the study is to introduce,test a.d verify a. a.va.ced ima.e processing technique w.th a.gorithms to a.toma.ica.ly extra.t tumour sections from bra.n MRI sca.s,fa.ilita.ing improved *** resea.ch intends to devise a.relia.le fra.ew.rk for detecting the BT region in the tw.dimensiona.(2D)MRI slice,a.d identifying its cla.s w.th improved *** methodology for the devised fra.ew.rk comprises the pha.es of:(i)Collection a.d resizing of ima.es,(ii)Implementa.ion a.d Segmenta.ion of Convolutiona. Neura. Netw.rk(CNN),(iii)Deep fea.ure extra.tion,(iv)Ha.dcra.ted fea.ure extra.tion,(v)Moth-Fla.e-a.gorithm(MFa.supported fea.ure reduction,a.d(vi)Performa.ce *** study utilized clinica.-gra.e bra.n MRI of BRa.S a.d TCIa.da.a.ets for the *** fra.ew.rk segments detected the glioma.low.high gra.e)a.d gliobla.toma.cla.s *** w.rk helped to get a.segmenta.ion a.cura.y of over 98%w.th VGG-UNet a.d a.cla.sifica.ion a.cura.y of over 98%w.th the VGG16 *** study ha. confirmed tha. the implemented fra.ew.rk is very efficient in detecting the BT in MRI slices w.th/w.thout the skull section.

关键词： Bra.n tumour VGG-UNet VGG16 moth-fla.e-a.gorithm cla.sifica.ion

Refined a.a.-Net:Lightw.ight Deep Lea.ning Model for Improved Segmenta.ion Performa.ce of Optic Cup a.d Disc for Gla.coma.Dia.nosis

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Ma.eria.s & Continua 2024年第7期80卷 1381-1405页

作者： Khursheed a.ra.gzeb syed Irta.a.Ha.der Musa.d a.hussein Depa.tment of Computer Engineering College of Computer and Information SciencesKing Saud UniversityP.O.Box 51178Riyadh11543Saudi Arabia

In this w.rk,w. a.m to introduce some modifica.ions to the a.a.-Net deep neura. netw.rk(DNN)model for segmenting optic cup(OC)a.d optic disc(OD)in retina. fundus ima.es to estima.e the cup-to-disc ra.io(CDR).The CDR is a.relia.le mea.ure for the ea.ly dia.nosis of *** this study,w. developed a.lightw.ight DNN model for OC a.d OD segmenta.ion in retina. fundus *** DNN model is ba.ed on modifica.ions to a.a.-Net,incorpora.ing a. a.a.orphic depth embedding *** reduce computa.iona. complexity,w. employ a.fixed filter size for a.l convolution la.ers in the encoder a.d decoder sta.es a. the netw.rk *** modifica.ion significa.tly reduces the number of tra.na.le pa.a.eters,ma.ing the model lightw.ight a.d suita.le for resource-constra.ned *** eva.ua.e the performa.ce of the developed model using tw. publicly a.a.la.le retina. ima.e da.a.a.es,na.ely RIM-ONE a.d *** results demonstra.e promising OC segmenta.ion performa.ce a.ross most sta.da.d eva.ua.ion metrics w.ile a.hieving a.a.ogous results for OD *** used tw. retina. fundus ima.e da.a.a.es na.ed RIM-ONE a.d Drishti-GS tha. conta.ned 159 ima.es a.d 101 retina. ima.es,*** OD segmenta.ion using the RIM-ONE w. obta.n a. f1-score(F1),Ja.ca.d coefficient(JC),a.d overla.ping error(OE)of 0.950,0.9219,a.d 0.0781,***,for OC segmenta.ion using the sa.e da.a.a.es,w. a.hieve scores of 0.8481(F1),0.7428(JC),a.d 0.2572(OE).Ba.ed on these experimenta. results a.d the significa.tly low.r number of tra.na.le pa.a.eters,w. conclude tha. the developed model is highly suita.le for the ea.ly dia.nosis of gla.coma.by a.cura.ely estima.ing the CDR.

关键词： Refined a.a.-Net pa.a.eter tuning deep lea.ning optic cup optic disc cup-to-disc ra.io gla.coma.dia.nosis

a. efficient permuta.ion a.proa.h for SbPN-ba.ed symmetric block ciphers

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Cybersecurity 2024年第2期7卷 113-130页

作者： Mir Na.ish M.Ta.iq Ba.da. Insha.syed Sheena.Ba.da. Depa.tment of Electronics a.d Instrumenta.ion Technology University ofKashmirSrinagar 190006India

It is cha.lenging to devise lightw.ight cryptogra.hic primitives efficient in both ha.dw.re a.d softw.re tha. ca. provide a. optimum level of security to diverse Internet of Things a.plica.ions running on low.end constra.ned ***,a. effcient ha.dw.re design a.proa.h tha. requires some specific ha.dw.re resource ma. not be effcient if implemented in *** bit Permuta.ion Netw.rk ba.ed ciphers such a. PRESENT a.d GIFT a.e effcient,lightw.ight cryptogra.hic ha.dw.re design *** ciphers introduce confusion a.d diffu-sion by employing a.4×4 sta.ic substitution box a.d bit *** bit-w.se permuta.ion is rea.ised by sim-ple rerouting,w.ich is most cost-effective to implement in ha.dw.re,resulting in negligible pow.r ***,this method is highly resource-consuming in softw.re,pa.ticula.ly for la.ge block-sized ciphers,w.th ea.h single-bit permuta.ion requiring multiple *** pa.er proposes a.novel softw.re-ba.ed design a.proa.h for permuta.ion opera.ion in Substitution bit Permuta.ion Netw.rk ba.ed ciphers using a.bit-ba.ding *** conventiona. permuta.ion using bit rota.ion a.d the proposed a.proa.h ha.e been implemented,a.a.ysed a.d compa.ed for GIFT a.d PRESENT ciphers on a.M Cortex-M3-ba.ed LPC1768 development pla.form w.th KEIL MDK used a. a. Integra.ed Development *** rea.-time performa.ce compa.ison betw.en conven-tiona. a.d the proposed a.proa.hes in terms of memory(Ra./ROM)footprint,pow.r,energy a.d execution time ha. been ca.ried out using ULINKpro a.d ULINKplus debug a.a.ters for va.ious code a.d speed optimisa.ion *** proposed a.proa.h substa.tia.ly reduces execution time,energy a.d pow.r consumption for both PRE-SENT a.d GIFT ciphers,thus demonstra.ing the effciency of the proposed method for Substitution bit Permuta.ion Netw.rk ba.ed symmetric block ciphers.

关键词： Lightw.ight cryptogra.hy PRESENT GIFT SbPN Cortex-M LPC1768 Bit-ba.d memory

Conserva.ion la.0},Lie symmetries,self a.jointness,a.d soliton solutions for the Selkov–Schna.enberg system

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Communica.ions in Theoretica. Physics 2024年第2期76卷 29-44页

作者： Ka.hif a.i a.y R Sea.a.0} syed T R Rizvi Noor a.iz Depa.tment of Ma.hema.ics COMSATS University IslamabadLahore CampusPakistan Ma.hema.ics Depa.tment Faculty of ScienceTaibah UniversityAl-Madinah Al-Munawarah41411Saudi Arabia

In this a.ticle,w. explore the fa.ous Selkov–Schna.enberg(SS)system of coupled nonlinea. pa.tia. differentia. equa.ions(PDEs)for Lie symmetry a.a.ysis,self-a.jointness,a.d conserva.ion ***,miscella.eous soliton solutions like da.k,bright,periodic,ra.iona.,Ja.obia. elliptic function,w.ierstra.s elliptic function,a.d hyperbolic solutions of the SS system w.ll be a.hieved by a.w.ll-know. technique ca.led sub-ordina.y differentia. *** these results a.e displa.ed gra.hica.ly by 3D,2D,a.d contour plots.

关键词： Selkov-Schna.enberg system Lie symmetry a.a.ysis conserva.ion la.0} a.jointness integra.ility

Detection of onset of a.glomera.ion in a.bubbling fluidized bed bioma.s combustor using rea.tive Euleria. computa.iona. fluid dyna.ics

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Pa.ticuology 2024年第7期90卷 504-515页

作者： a.dulla. Ta.leem a.ta.Ulla. Fei Li Qun Yi w.llia. Nimmo syed Shera. Da.od Depa.tment of Chemica. Engineering Pakistan Institute of Engineering and Applied Sciences(PIEAS)Lehtrar RoadP.O.Nilore45650IslamabadPakistan Institute of Process Engineering Chinese Academy of SciencesBeijing100190China School of Chemica. Engineering a.d Pha.ma.y Wuhan Institute of TechnologyWuhan430205China Energy Engineering Group Energy 2050Department of Mechanical EngineeringUniversity of SheffieldSheffieldS37RDUnited Kingdom Institute of Energy a.d Environmenta. Engineering Faculty of ElectricalEnergy&Environmental EngineeringUniversity of the PunjabQuaid-e-Azam CampusLahorePakistan Energy Engineering Resea.ch a.d Development Centre Faculty of ElectricalEnergy&Environmental EngineeringUniversity of the PunjabQuaid-e-Azam CampusLahorePakistan

The choice of a.type of combustion technology to be used for hea. or pow.r genera.ion depends on economic,technica.,opera.iona. a.d fuel a.a.la.ility *** benefits a.socia.ed w.th the evolving ma.ket driven by the fluidised bed combustion(FBC)technology ca.not be overlooked especia.ly w.en ga.ged a. 65 Gw.h of w.rldw.de insta.led ca.a.ity a.ongside a.ded benefits of ha.dling fuel va.ia.ion,low.polluta.t emissions a.d high combustion *** or bioma.s w.ste w.ll continue to ha.e a.vita. role to pla. in the future FBC technology-ba.ed pow.r *** often conta.ns high levels of inorga.ic species tha. ca. form sticky a.glomera.es posing a.significa.t risk to boiler opera.ion resulting in unscheduled *** a.ded complexity of the beha.iour of the fuel a.d bed ma.eria. mix highlights the requirement for simula.ion models to identify a.glomera.ion to help improve the overa.l performa.ce a.d relia.ility of FBC *** resolve this problem,this resea.ch devised a.simula.ion stra.egy for the detection of a.glomera.ion using the Euleria.–Euleria. *** developed modelling stra.egy is va.ida.ed w.th the experimenta. da.a.a.a.la.le in litera.ure for tw.-dimensiona. simplified geometry of a.pilot-sca.e fluidised bed *** model results w.re found promising a.d robust to predict bed defluidisa.ion times a.d other pa.a.eters consistent w.th the experimenta. da.a.

关键词： Fluidized bed combustion Euleria. Models Bioma.s a.glomera.ion Defluidisa.ion

Enha.cing Skin Ca.cer Dia.nosis w.th Deep Lea.ning:a.Hybrid CNN-RNN a.proa.h

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Ma.eria.s & Continua 2024年第4期79卷 1497-1519页

作者： syeda.Sha.a.la.Za.een Gua.gmin Sun Ma.w.sh Kundi syed Furqa. Qa.ri Sa.ma. Qa.ri Fa.ulty of Informa.ion Technology Beijing University of TechnologyBeijing100124China Computer Science Interna.iona. Engineering Colla.e Maynooth UniversityKildareW23 F2H6Irland Resea.ch Center for Da.a.Hub a.d Security Zhejiang LabHangzhou311121China Computer Science Depa.tment MNS University of AgricultureMultan59220Pakistan

Skin ca.cer dia.nosis is difficult due to lesion presenta.ion va.ia.ility. Conventiona.methods struggle to ma.ua.lyextra.t fea.ures a.d ca.ture lesions spa.ia. a.d tempora. va.ia.ions. This study introduces a.deep lea.ning-ba.edConvolutiona. a.d Recurrent Neura. Netw.rk (CNN-RNN) model w.th a.ResNet-50 a.chitecture w.ich useda. the fea.ure extra.tor to enha.ce skin ca.cer cla.sifica.ion. Levera.ing synergistic spa.ia. fea.ure extra.tiona.d tempora. sequence lea.ning, the model demonstra.es robust performa.ce on a.da.a.et of 9000 skin lesionphotos from nine ca.cer types. Using pre-tra.ned ResNet-50 for spa.ia. da.a.extra.tion a.d Long Short-TermMemory (LSTM) for tempora. dependencies, the model a.hieves a.high a.era.e recognition a.cura.y, surpa.singprevious methods. The comprehensive eva.ua.ion, including a.cura.y, precision, reca.l, a.d F1-score, underscoresthe model’s competence in ca.egorizing skin ca.cer types. This resea.ch contributes a.sophistica.ed model a.dva.ua.le guida.ce for deep lea.ning-ba.ed dia.nostics, a.so this model excels in overcoming spa.ia. a.d tempora.complexities, offering a.sophistica.ed solution for derma.ologica. dia.nostics resea.ch.

关键词： Skin ca.cer cla.sifica.ion deep lea.ning Convolutiona. Neura. Netw.rk(CNN) RNN ResNet-50

Ena.ling Efficient Da.a.Tra.smission in w.reless Sensor Netw.rks-Ba.ed IoT a.plica.ion

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Ma.eria.s & Continua 2024年第6期79卷 4197-4218页

作者： Ibra.eem a.-Hejri Fa.a. a.zedin Sulta. a.muha.ma.i Na.em Firdous syed Informa.ion a.d Computer Science Depa.tment King Fahd University of Petroleum andMineralsDhahran31261Saudi Arabia Center for Cyber Resilience a.d Trust(CREST) Deakin UniversityGeelong3220Australia

The use of the Internet of Things(IoT)is expa.ding a. a. unprecedented sca.e in ma.y critica. a.plica.ions due to the a.ility to interconnect a.d utilize a.plethora.of w.de ra.ge of *** critica. infra.tructure doma.ns like oil a.d ga. supply,intelligent tra.sporta.ion,pow.r grids,a.d a.tonomous a.riculture,it is essentia. to gua.a.tee the confidentia.ity,integrity,a.d a.thenticity of da.a.collected a.d ***,the limited resources coupled w.th the heterogeneity of IoT devices ma.e it inefficient or sometimes infea.ible to a.hieve secure da.a.tra.smission using tra.itiona. cryptogra.hic ***,designing a.lightw.ight secure da.a.tra.smission scheme is becoming *** this a.ticle,w. propose lightw.ight secure da.a.tra.smission(LSDT)scheme for IoT *** consists of three pha.es a.d utilizes a. effective combina.ion of symmetric keys a.d the Elliptic Curve Menezes-Qu-Va.stone a.ymmetric key a.reement *** design the simula.ion environment a.d experiments to eva.ua.e the performa.ce of the LSDT scheme in terms of communica.ion a.d computa.ion *** a.d performa.ce a.a.ysis indica.es tha. the LSDT scheme is secure,suita.le for IoT a.plica.ions,a.d performs better in compa.ison to other rela.ed security schemes.

关键词： IoT lightw.ight computa.ion complexity communica.ion overhea. cybersecurity threa.s threa. prevention secure da.a.tra.smission w.reless Sensor Netw.rks(w.Ns) elliptic curve cryptogra.hy

Model a.nostic Meta.Lea.ning(Ma.L)-Ba.ed Ensemble Model for a.cura.e Detection of w.ea. Disea.es Using Vision Tra.sformer a.d Gra.h Neura. Netw.rks

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Ma.eria.s & Continua 2024年第5期79卷 2795-2811页

作者： Ya.ir Ma.sood syed Muha.ma. Usma. Musa.d a.hussein Khursheed a.ra.gzeb Shehza. Kha.id Muha.ma. Zuba.r Depa.tment of Crea.ive Technologies Faculty of Computing and Artificial IntelligenceAir UniversityIslamabad44000Pakistan Depa.tment of Computer Engineering College of Computer and Information SciencesKing Saud UniversityP.O.Box 51178Riyadh11543Saudi Arabia Depa.tment of Computer Engineering Bahria School of Engineering and Applied SciencesBahria UniversityIslamabad44000Pakistan La.ora.ory for Neuro&Psychophysiology KU LeuvenLeuven3000Belgium

w.ea. is a.critica. crop,extensively consumed w.rldw.de,a.d its production enha.cement is essentia. to meet esca.a.ing *** presence of disea.es like stem rust,lea. rust,yellow.rust,a.d ta. spot significa.tly diminishes w.ea. yield,ma.ing the ea.ly a.d precise identifica.ion of these disea.es vita. for effective disea.e *** a.va.cements in deep lea.ning a.gorithms,resea.chers ha.e proposed ma.y methods for the a.toma.ed detection of disea.e pa.hogens;how.ver,a.cura.ely detectingmultiple disea.e pa.hogens simulta.eously rema.ns a.*** cha.lenge a.ises due to the sca.city of RGB ima.es for multiple disea.es,cla.s imba.a.ce in existing public da.a.ets,a.d the difficulty in extra.ting fea.ures tha. discrimina.e betw.en multiple cla.ses of disea.e *** this resea.ch,a.novel method is proposed ba.ed on Tra.sfer Genera.ive a.versa.ia. Netw.rks for a.gmenting existing da.a.thereby overcoming the problems of cla.s imba.a.ce a.d da.a.*** study proposes a.customized a.chitecture of Vision Tra.sformers(ViT),w.ere the fea.ure vector is obta.ned by conca.ena.ing fea.ures extra.ted from the custom ViT a.d Gra.h Neura. *** pa.er a.so proposes a.Model a.nosticMeta.Lea.ning(Ma.L)ba.ed ensemble cla.sifier for a.cura.e *** proposedmodel,va.ida.ed on public da.a.ets for w.ea. disea.e pa.hogen cla.sifica.ion,a.hieved a.test a.cura.y of 99.20%a.d a. F1-score of 97.95%.Compa.ed w.th existing sta.e-of-the-a.t methods,this proposed model outperforms in terms of a.cura.y,F1-score,a.d the number of disea.e pa.hogens *** future,more disea.es ca. be included for detection a.ong w.th some other moda.ities like pests a.d w.ed.

关键词： w.ea. disea.e detection deep lea.ning vision tra.sformer gra.h neura. netw.rk model a.nostic meta.lea.ning

Deep Lea.ning a.proa.h for Ha.d Gesture Recognition:a.plica.ions in Dea. Communica.ion a.d Hea.thca.e

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Ma.eria.s & Continua 2024年第1期78卷 127-144页

作者： Khursheed a.ra.gzeb Kha.id Ja.eed Musa.d a.hussein Ima. Rida.syed Irta.a.Ha.der a.ubha.Pa.a.ha. Depa.tment of Computer Engineering College of Computer and Information SciencesKing Saud UniversityP.O.Box 51178Riyadh11543Kingdom of Saudi Arabia Depa.tment of Computer Engineering College of Computing and InformaticsUniversity of SharjahSharjah27272United Arab Emirates La.ora.ory Biomecha.ics a.d Bioengineering University of Technology of CompiegneCompiegne60200France Depa.tment of Computer Science a.d Engineering Manipal University JaipurJaipur303007India

Ha.d gestures ha.e been used a. a.significa.t mode of communica.ion since the a.vent of huma. *** fa.ilita.ing huma.-computer intera.tion(HCI),ha.d gesture recognition(HGRoc)technology is crucia. for sea.less a.d error-free *** technology is pivota. in hea.thca.e a.d communica.ion for the dea. *** significa.t a.va.cements in computer vision-ba.ed gesture recognition for la.gua.e understa.ding,tw. considera.le cha.lenges persist in this field:(a.limited a.d common gestures a.e considered,(b)processing multiple cha.nels of informa.ion a.ross a.netw.rk ta.es huge computa.iona. time during discrimina.ive fea.ure ***,a.novel ha.d vision-ba.ed convolutiona. neura. netw.rk(CNN)model na.ed(HVCNNM)offers severa. benefits,nota.ly enha.ced a.cura.y,robustness to va.ia.ions,rea.-time performa.ce,reduced cha.nels,a.d ***,these models ca. be optimized for rea.-time performa.ce,lea.n from la.ge a.ounts of da.a.a.d a.e sca.a.le to ha.dle complex recognition ta.ks for efficient huma.-computer *** proposed model w.s eva.ua.ed on tw. cha.lenging da.a.ets,na.ely the Ma.sey University Da.a.et(MUD)a.d the a.erica. Sign La.gua.e(a.L)a.pha.et Da.a.et(a.La.).On the MUD a.d a.La. da.a.ets,HVCNNM a.hieved a.score of 99.23% a.d 99.00%,*** results demonstra.e the effectiveness of CNN a. a.promising HGRoc *** findings suggest tha. the proposed model ha.e potentia. roles in a.plica.ions such a. sign la.gua.e recognition,huma.-computer intera.tion,a.d robotics.

关键词： Computer vision deep lea.ning ga.t recognition sign la.gua.e recognition ma.hine lea.ning