Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais
Ano de defesa: | 2019 |
---|---|
Autor(a) principal: | |
Orientador(a): | |
Banca de defesa: | |
Tipo de documento: | Dissertação |
Tipo de acesso: | Acesso aberto |
Idioma: | por |
Instituição de defesa: |
Universidade Estadual de Feira de Santana
|
Programa de Pós-Graduação: |
Mestrado em Computa??o Aplicada
|
Departamento: |
DEPARTAMENTO DE CI?NCIAS EXATAS
|
País: |
Brasil
|
Palavras-chave em Português: | |
Palavras-chave em Inglês: | |
Área do conhecimento CNPq: | |
Link de acesso: | http://tede2.uefs.br:8080/handle/tede/1342 |
Resumo: | Glomerulopathies, kidney diseases, affect thousands of people in Brazil and in the entire world, this number is growing constantly. The glomeruli are microscopic structures present in kidney and your examination by a doctor determines the kind and the degree of kidney disease. Kidney tissue images can be scanned or photographed, enabling the computational processing. Nowadays, detection and segmentation are made manually by a pathologist doctor. Thus, this research aims at propose a glomeruli automatic detection method on histological digital kidney tissue images. For this, we use deep learning techniques to train capable models to automate this task. Digital images photographed in varied approximation scales was used to compose train and test datasets. Tensorflow Object Detection API (Application Programming Interface) framework was used implements, train and test the models SSD (Single Shot Detection) Inception V2(SI2) and Faster RCNN (Region-based Convolutional Neural Network) Inception V2 (FRI2). Reaching 0.8831 mAP - 0.94 F1 Score when using the SI2 model, 0.8723 mAP and 0.97 F1 Score when utilizing FRI2 model. The SI2 model. The SI2 model is the most efficient for this task because it is 64% faster in training time and 98% faster in detecting glomeruli in each image. This work demonstrate the efficiency of deep learning techniques as solution for this problem, advancing the improvement of techniques for gloeruli automated detection. |
id |
UEFS_48495710734ac32b8906314eb178d789 |
---|---|
oai_identifier_str |
oai:tede2.uefs.br:8080:tede/1342 |
network_acronym_str |
UEFS |
network_name_str |
Biblioteca Digital de Teses e Dissertações da UEFS |
repository_id_str |
|
spelling |
Angelo, Michele F?lvia26554093885http://lattes.cnpq.br/603227384984728500667602593http://lattes.cnpq.br/5711710989208932Rehem, Jonathan Moreira Cardozo2022-04-19T20:45:49Z2019-08-19REHEM, Jonathan Moreira Cardozo. Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais. 2019. 95 f. Disserta??o (Mestrado em Computa??o Aplicada)- Universidade Estadual de Feira de Santana, Feira de Santana, 2019.http://tede2.uefs.br:8080/handle/tede/1342Glomerulopathies, kidney diseases, affect thousands of people in Brazil and in the entire world, this number is growing constantly. The glomeruli are microscopic structures present in kidney and your examination by a doctor determines the kind and the degree of kidney disease. Kidney tissue images can be scanned or photographed, enabling the computational processing. Nowadays, detection and segmentation are made manually by a pathologist doctor. Thus, this research aims at propose a glomeruli automatic detection method on histological digital kidney tissue images. For this, we use deep learning techniques to train capable models to automate this task. Digital images photographed in varied approximation scales was used to compose train and test datasets. Tensorflow Object Detection API (Application Programming Interface) framework was used implements, train and test the models SSD (Single Shot Detection) Inception V2(SI2) and Faster RCNN (Region-based Convolutional Neural Network) Inception V2 (FRI2). Reaching 0.8831 mAP - 0.94 F1 Score when using the SI2 model, 0.8723 mAP and 0.97 F1 Score when utilizing FRI2 model. The SI2 model. The SI2 model is the most efficient for this task because it is 64% faster in training time and 98% faster in detecting glomeruli in each image. This work demonstrate the efficiency of deep learning techniques as solution for this problem, advancing the improvement of techniques for gloeruli automated detection.As glomerulopatias, doen?as renais, acometem milhares de pessoas no Brasil e no mundo e este n?mero vem crescendo. Os glom?rulos s?o estruturas microsc?picas presentes nos rins e sua an?lise por um m?dico patologista ? o que determina o tipo e o grau da doen?a renal. Imagens dos tecidos renais podem ser digitalizadas ou fotografadas, o que torna poss?vel o processamento por computador. Atualmente, a detec??o e a separa??o de glom?rulos ? feita manualmente pelo patologista. Assim, esta pesquisa tem como objetivo propor um m?todo de detec??o autom?tico de glom?rulos em imagens histol?gicas renais digitais. Para isso, foram utilizadas t?cnicas de aprendizagem profunda a fim de treinar modelos que fossem capazes de automatizar esta tarefa. Imagens digitais de l?minas histol?gicas fotografadas em variadas escalas de aproxima??o foram utilizadas para compor os datasets de treinamento e testes. O framework Tensorflow Object Detection API foi utilizado como plataforma de implementa??o no treinamento e testes dos modelos SSD Inception V2 e Faster RCNN Inception V2. Obteve-se 0.8831 mAP e 0.94 F1 Score utilizando o modelo SI2, e 0.8723 mAP e 0.97 F1 Score utilizando o modelo FRI2. O modelo SI2 ? o mais eficiente para esta tarefa, j? que ? 64% mais r?pido no tempo necess?rio para o treinamento e 98% mais r?pido na detec??o de glom?rulos em cada imagem. Este trabalho demonstra a efici?ncia do Deep Learning na resolu??o deste problema, avan?ando no aperfei?oamento das t?cnicas de detec??o autom?tica de glom?rulos.Submitted by Ricardo Cedraz Duque Moliterno (ricardo.moliterno@uefs.br) on 2022-04-19T20:45:49Z No. of bitstreams: 1 dissertacao_jonathan_cardozo_versao_final.pdf: 4815629 bytes, checksum: 0f1692645a1026115d04d06bef64055a (MD5)Made available in DSpace on 2022-04-19T20:45:49Z (GMT). No. of bitstreams: 1 dissertacao_jonathan_cardozo_versao_final.pdf: 4815629 bytes, checksum: 0f1692645a1026115d04d06bef64055a (MD5) Previous issue date: 2019-08-19application/pdfporUniversidade Estadual de Feira de SantanaMestrado em Computa??o AplicadaUEFSBrasilDEPARTAMENTO DE CI?NCIAS EXATASDeep learningRedes Neurais ConvolucionaisGlom?rulosDetec??o de objetosDeep learningConvolutional Neural NetworksGlomeruliObject DetectionCIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAODetec??o autom?tica de glom?rulos em imagens histol?gicas renais digitaisinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesis303317282311144204600600600-54868328166115062113671711205811204509info:eu-repo/semantics/openAccessreponame:Biblioteca Digital de Teses e Dissertações da UEFSinstname:Universidade Estadual de Feira de Santana (UEFS)instacron:UEFSORIGINALdissertacao_jonathan_cardozo_versao_final.pdfdissertacao_jonathan_cardozo_versao_final.pdfapplication/pdf4815629http://tede2.uefs.br:8080/bitstream/tede/1342/2/dissertacao_jonathan_cardozo_versao_final.pdf0f1692645a1026115d04d06bef64055aMD52LICENSElicense.txtlicense.txttext/plain; charset=utf-82089http://tede2.uefs.br:8080/bitstream/tede/1342/1/license.txt7b5ba3d2445355f386edab96125d42b7MD51tede/13422022-04-19 17:45:49.067oai:tede2.uefs.br:8080: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Biblioteca Digital de Teses e Dissertaçõeshttp://tede2.uefs.br:8080/PUBhttp://tede2.uefs.br:8080/oai/requestbcuefs@uefs.br|| bcref@uefs.br||bcuefs@uefs.bropendoar:2022-04-19T20:45:49Biblioteca Digital de Teses e Dissertações da UEFS - Universidade Estadual de Feira de Santana (UEFS)false |
dc.title.por.fl_str_mv |
Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais |
title |
Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais |
spellingShingle |
Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais Rehem, Jonathan Moreira Cardozo Deep learning Redes Neurais Convolucionais Glom?rulos Detec??o de objetos Deep learning Convolutional Neural Networks Glomeruli Object Detection CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO |
title_short |
Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais |
title_full |
Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais |
title_fullStr |
Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais |
title_full_unstemmed |
Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais |
title_sort |
Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais |
author |
Rehem, Jonathan Moreira Cardozo |
author_facet |
Rehem, Jonathan Moreira Cardozo |
author_role |
author |
dc.contributor.advisor1.fl_str_mv |
Angelo, Michele F?lvia |
dc.contributor.advisor1ID.fl_str_mv |
26554093885 |
dc.contributor.advisor1Lattes.fl_str_mv |
http://lattes.cnpq.br/6032273849847285 |
dc.contributor.authorID.fl_str_mv |
00667602593 |
dc.contributor.authorLattes.fl_str_mv |
http://lattes.cnpq.br/5711710989208932 |
dc.contributor.author.fl_str_mv |
Rehem, Jonathan Moreira Cardozo |
contributor_str_mv |
Angelo, Michele F?lvia |
dc.subject.por.fl_str_mv |
Deep learning Redes Neurais Convolucionais Glom?rulos Detec??o de objetos |
topic |
Deep learning Redes Neurais Convolucionais Glom?rulos Detec??o de objetos Deep learning Convolutional Neural Networks Glomeruli Object Detection CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO |
dc.subject.eng.fl_str_mv |
Deep learning Convolutional Neural Networks Glomeruli Object Detection |
dc.subject.cnpq.fl_str_mv |
CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO |
description |
Glomerulopathies, kidney diseases, affect thousands of people in Brazil and in the entire world, this number is growing constantly. The glomeruli are microscopic structures present in kidney and your examination by a doctor determines the kind and the degree of kidney disease. Kidney tissue images can be scanned or photographed, enabling the computational processing. Nowadays, detection and segmentation are made manually by a pathologist doctor. Thus, this research aims at propose a glomeruli automatic detection method on histological digital kidney tissue images. For this, we use deep learning techniques to train capable models to automate this task. Digital images photographed in varied approximation scales was used to compose train and test datasets. Tensorflow Object Detection API (Application Programming Interface) framework was used implements, train and test the models SSD (Single Shot Detection) Inception V2(SI2) and Faster RCNN (Region-based Convolutional Neural Network) Inception V2 (FRI2). Reaching 0.8831 mAP - 0.94 F1 Score when using the SI2 model, 0.8723 mAP and 0.97 F1 Score when utilizing FRI2 model. The SI2 model. The SI2 model is the most efficient for this task because it is 64% faster in training time and 98% faster in detecting glomeruli in each image. This work demonstrate the efficiency of deep learning techniques as solution for this problem, advancing the improvement of techniques for gloeruli automated detection. |
publishDate |
2019 |
dc.date.issued.fl_str_mv |
2019-08-19 |
dc.date.accessioned.fl_str_mv |
2022-04-19T20:45:49Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/masterThesis |
format |
masterThesis |
status_str |
publishedVersion |
dc.identifier.citation.fl_str_mv |
REHEM, Jonathan Moreira Cardozo. Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais. 2019. 95 f. Disserta??o (Mestrado em Computa??o Aplicada)- Universidade Estadual de Feira de Santana, Feira de Santana, 2019. |
dc.identifier.uri.fl_str_mv |
http://tede2.uefs.br:8080/handle/tede/1342 |
identifier_str_mv |
REHEM, Jonathan Moreira Cardozo. Detec??o autom?tica de glom?rulos em imagens histol?gicas renais digitais. 2019. 95 f. Disserta??o (Mestrado em Computa??o Aplicada)- Universidade Estadual de Feira de Santana, Feira de Santana, 2019. |
url |
http://tede2.uefs.br:8080/handle/tede/1342 |
dc.language.iso.fl_str_mv |
por |
language |
por |
dc.relation.program.fl_str_mv |
303317282311144204 |
dc.relation.confidence.fl_str_mv |
600 600 600 |
dc.relation.department.fl_str_mv |
-5486832816611506211 |
dc.relation.cnpq.fl_str_mv |
3671711205811204509 |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.publisher.none.fl_str_mv |
Universidade Estadual de Feira de Santana |
dc.publisher.program.fl_str_mv |
Mestrado em Computa??o Aplicada |
dc.publisher.initials.fl_str_mv |
UEFS |
dc.publisher.country.fl_str_mv |
Brasil |
dc.publisher.department.fl_str_mv |
DEPARTAMENTO DE CI?NCIAS EXATAS |
publisher.none.fl_str_mv |
Universidade Estadual de Feira de Santana |
dc.source.none.fl_str_mv |
reponame:Biblioteca Digital de Teses e Dissertações da UEFS instname:Universidade Estadual de Feira de Santana (UEFS) instacron:UEFS |
instname_str |
Universidade Estadual de Feira de Santana (UEFS) |
instacron_str |
UEFS |
institution |
UEFS |
reponame_str |
Biblioteca Digital de Teses e Dissertações da UEFS |
collection |
Biblioteca Digital de Teses e Dissertações da UEFS |
bitstream.url.fl_str_mv |
http://tede2.uefs.br:8080/bitstream/tede/1342/2/dissertacao_jonathan_cardozo_versao_final.pdf http://tede2.uefs.br:8080/bitstream/tede/1342/1/license.txt |
bitstream.checksum.fl_str_mv |
0f1692645a1026115d04d06bef64055a 7b5ba3d2445355f386edab96125d42b7 |
bitstream.checksumAlgorithm.fl_str_mv |
MD5 MD5 |
repository.name.fl_str_mv |
Biblioteca Digital de Teses e Dissertações da UEFS - Universidade Estadual de Feira de Santana (UEFS) |
repository.mail.fl_str_mv |
bcuefs@uefs.br|| bcref@uefs.br||bcuefs@uefs.br |
_version_ |
1796793644394479616 |