| |
|
|
 | Acesso ao texto completo restrito à biblioteca da Embrapa Gado de Leite. Para informações adicionais entre em contato com cnpgl.biblioteca@embrapa.br. |
|
Registro Completo |
|
Biblioteca(s): |
Embrapa Gado de Leite. |
|
Data corrente: |
20/03/2024 |
|
Data da última atualização: |
20/03/2024 |
|
Tipo da produção científica: |
Artigo em Periódico Indexado |
|
Autoria: |
OLIVEIRA, D. A. B.; BRESOLIN, T.; COELHO, S. G.; CAMPOS, M. M.; LAGE, C. F. A.; LEÃO, J. M.; PEREIRA, L. G. R.; HERNANDEZ, L.; DOREA, J. R. R. |
|
Afiliação: |
DARIO A. B. OLIVEIRA, UNIVERSITY OF WISCONSIN-MADISON; TIAGO BRESOLIN, UNIVERSITY OF ILLINOIS; SANDRA G. COELHO, UNIVERSIDADE FEDERAL DE MINAS GERAIS; MARIANA MAGALHAES CAMPOS, CNPGL; UNIVERSIDADE FEDERAL DE MINAS GERAIS; UNIVERSIDADE FEDERAL DE MINAS GERAIS; LUIZ GUSTAVO RIBEIRO PEREIRA, CNPGL; LAURA HERNANDEZ, UNIVERSITY OF WISCONSIN-MADISON; JOÃO R. R. DOREA, UNIVERSITY OF WISCONSIN-MADISON. |
|
Título: |
A polar transformation augmentation approach for enhancing mammary gland segmentation in ultrasound images. |
|
Ano de publicação: |
2024 |
|
Fonte/Imprenta: |
Computers and Electronics in Agriculture, v. 220, 108825, 2024. |
|
DOI: |
https://doi.org/10.1016/j.compag.2024.108825 |
|
Idioma: |
Inglês |
|
Conteúdo: |
Environmental factors can detrimentally affect mammary gland development, leading to negative impacts on milk secretion in mammals. Ultrasonography serves as a non-invasive and non-destructive method for assessing mammary gland characteristics and development. Deep learning approaches enable automated monitoring of mammary gland development, though they typically require large, labeled datasets that may be limited by data collection constraints. This study aimed to develop and evaluate a polar transformation-based augmentation strategy to enhance the performance of deep learning algorithms for mammary gland segmentation in small datasets. We collected 405 ultrasound images of mammary glands (front and rear quarters) from 29 crossbred F1 Holstein x Gyr calves aged 5 to 11 weeks. The parenchyma tissue in these images was manually annotated using the VGG Image Annotator. A leave-one-animal-out cross-validation approach was employed to train the semantic segmentation algorithm. In this approach, all images from one calf were used as a testing set, and images from the remaining 28 calves were used for training in each of the 29 iterations. Our proposed method involved utilizing a polar transform technique for data augmentation in ultrasound images and the PSPNet deep learning algorithm for image segmentation. The average F1-score on the testing set was 54% in week 1, 70% in week 2, and 75% in week 3. Our findings revealed that the algorithm’s performance was suboptimal for images with very small parenchyma (week 1). However, as the mammary gland developed, the identification and segmentation of parenchymal tissue significantly improved. The performance of deep learning algorithms in segmenting small tissues could potentially be enhanced by using larger datasets and higher resolution images. In conclusion, our study demonstrates that polar transformation is an effective strategy for augmenting mammary gland ultrasound images, which in turn improves the performance of deep neural networks in segmenting parenchymal tissue. MenosEnvironmental factors can detrimentally affect mammary gland development, leading to negative impacts on milk secretion in mammals. Ultrasonography serves as a non-invasive and non-destructive method for assessing mammary gland characteristics and development. Deep learning approaches enable automated monitoring of mammary gland development, though they typically require large, labeled datasets that may be limited by data collection constraints. This study aimed to develop and evaluate a polar transformation-based augmentation strategy to enhance the performance of deep learning algorithms for mammary gland segmentation in small datasets. We collected 405 ultrasound images of mammary glands (front and rear quarters) from 29 crossbred F1 Holstein x Gyr calves aged 5 to 11 weeks. The parenchyma tissue in these images was manually annotated using the VGG Image Annotator. A leave-one-animal-out cross-validation approach was employed to train the semantic segmentation algorithm. In this approach, all images from one calf were used as a testing set, and images from the remaining 28 calves were used for training in each of the 29 iterations. Our proposed method involved utilizing a polar transform technique for data augmentation in ultrasound images and the PSPNet deep learning algorithm for image segmentation. The average F1-score on the testing set was 54% in week 1, 70% in week 2, and 75% in week 3. Our findings revealed that the algorithm’s performance was suboptimal for images... Mostrar Tudo |
|
Palavras-Chave: |
Polar transformation; Segmentação semântica; Semantic segmentation; Transformação polar; Ultrassonografia. |
|
Thesagro: |
Bovino; Glândula Mamaria; Ultrassom. |
|
Thesaurus Nal: |
Mammary glands; Ultrasonography. |
|
Categoria do assunto: |
L Ciência Animal e Produtos de Origem Animal |
|
Marc: |
LEADER 03118naa a2200349 a 4500
001 2163042
005 2024-03-20
008 2024 bl uuuu u00u1 u #d
024 7 $ahttps://doi.org/10.1016/j.compag.2024.108825$2DOI
100 1 $aOLIVEIRA, D. A. B.
245 $aA polar transformation augmentation approach for enhancing mammary gland segmentation in ultrasound images.$h[electronic resource]
260 $c2024
520 $aEnvironmental factors can detrimentally affect mammary gland development, leading to negative impacts on milk secretion in mammals. Ultrasonography serves as a non-invasive and non-destructive method for assessing mammary gland characteristics and development. Deep learning approaches enable automated monitoring of mammary gland development, though they typically require large, labeled datasets that may be limited by data collection constraints. This study aimed to develop and evaluate a polar transformation-based augmentation strategy to enhance the performance of deep learning algorithms for mammary gland segmentation in small datasets. We collected 405 ultrasound images of mammary glands (front and rear quarters) from 29 crossbred F1 Holstein x Gyr calves aged 5 to 11 weeks. The parenchyma tissue in these images was manually annotated using the VGG Image Annotator. A leave-one-animal-out cross-validation approach was employed to train the semantic segmentation algorithm. In this approach, all images from one calf were used as a testing set, and images from the remaining 28 calves were used for training in each of the 29 iterations. Our proposed method involved utilizing a polar transform technique for data augmentation in ultrasound images and the PSPNet deep learning algorithm for image segmentation. The average F1-score on the testing set was 54% in week 1, 70% in week 2, and 75% in week 3. Our findings revealed that the algorithm’s performance was suboptimal for images with very small parenchyma (week 1). However, as the mammary gland developed, the identification and segmentation of parenchymal tissue significantly improved. The performance of deep learning algorithms in segmenting small tissues could potentially be enhanced by using larger datasets and higher resolution images. In conclusion, our study demonstrates that polar transformation is an effective strategy for augmenting mammary gland ultrasound images, which in turn improves the performance of deep neural networks in segmenting parenchymal tissue.
650 $aMammary glands
650 $aUltrasonography
650 $aBovino
650 $aGlândula Mamaria
650 $aUltrassom
653 $aPolar transformation
653 $aSegmentação semântica
653 $aSemantic segmentation
653 $aTransformação polar
653 $aUltrassonografia
700 1 $aBRESOLIN, T.
700 1 $aCOELHO, S. G.
700 1 $aCAMPOS, M. M.
700 1 $aLAGE, C. F. A.
700 1 $aLEÃO, J. M.
700 1 $aPEREIRA, L. G. R.
700 1 $aHERNANDEZ, L.
700 1 $aDOREA, J. R. R.
773 $tComputers and Electronics in Agriculture$gv. 220, 108825, 2024.
Download
Esconder MarcMostrar Marc Completo |
|
Registro original: |
Embrapa Gado de Leite (CNPGL) |
|
|
Biblioteca |
ID |
Origem |
Tipo/Formato |
Classificação |
Cutter |
Registro |
Volume |
Status |
URL |
Voltar
|
|
|
| Registros recuperados : 618 | |
| Registros recuperados : 618 | |
|
| Nenhum registro encontrado para a expressão de busca informada. |
|
|
