Avaliação de abordagens semi-supervisionadas aplicadas a redes neurais convolucionais
Resumo
O presente trabalho foca em mitigar o problema geral de rotulação de amostras em bases de imagens utilizadas em redes neurais convolucionais. Para isso, foi aplicado o paradigma semi-supervisionado às redes neurais convolucionais. Foi proposta a comparação de duas técnicas semi-supervisionadas da literatura com este tipo de rede de aprendizado profundo e analizado o seu comportamento. Os experimentos, com três conjuntos de dados públicos, atestam que a agregação proposta obteve melhores resultados, ganhos de até 88% em acurácia, quando comparado com o paradigma supervisionado.
Palavras-chave:
Redes Neurais Convolucionais, aprendizado semi-supervisionado, aprendizado de máquina, auto-treinamento, co-treinamento
Referências
Amorim, W., de Rosa, G., Thomazella, R., Castanho, J. E., Dotto, F., Júnior, O., Marana, A., and Papa, J. (2019). Semi-supervised learning with connectivity-driven convolutional neural networks. Pattern Recognition Letters.
Chen, J., Feng, J., Sun, X., and Liu, Y. (2020). Co-training semi-supervised deep learning for sentiment classification of mooc forum posts. Symmetry, 12(1).
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database. In IEEE Conference on Computer Vision and Pattern Recognition, pages 248–255.
Elshennawy, N. M. and Ibrahim, D. M. (2020). Deep-pneumonia framework using deep learning models based on chest x-ray images. Diagnostics, 10(9).
Oliveira, P., de Carvalho Scabora, L., Cazzolato, M., Bedo, M., Traina, A., and Jr, C. (2017). Mammoset: An enhanced dataset of mammograms. In Dataset Showcase Workshop - DSW at the Brazilian Symposium on Databases, pages 1–11.
Orenstein, E., Beijbom, O., Peacock, E., and Sosik, H. (2015). Whoi-plankton-a large scale fine grained visual recognition benchmark dataset for plankton classification. Tech Report, pages 1–2.
Rosenberg, C., Hebert, M., and Schneiderman, H. (2005). Semi-supervised self-training of object detection models. In 2005 Seventh IEEE Workshops on Applications of Computer Vision - Volume 1, volume 1, pages 29–36.
T.R., S., Thomas, S. A., Kamath, V., and B., N. K. (2019). Hybrid xception model for human protein atlas image classification. In IEEE 16th India Council International Conference, pages 1–4.
van Engelen, J. E. and Hoos, H. H. (2019). A survey on semi-supervised learning. Machine Learning, 109:373–440.
Zhang, L., Lu, L., Nogues, I., Summers, R. M., Liu, S., and Yao, J. (2017). Deeppap: Deep convolutional networks for cervical cell classification. IEEE Journal of Biomedical and Health Informatics, 21:1633–1643.
Zhuang, F., Qi, Z., Duan, K., Xi, D., Zhu, Y., Zhu, H., Xiong, H., and He, Q. (2019). A comprehensive survey on transfer learning. CoRR, abs/1911.02685.
Chen, J., Feng, J., Sun, X., and Liu, Y. (2020). Co-training semi-supervised deep learning for sentiment classification of mooc forum posts. Symmetry, 12(1).
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database. In IEEE Conference on Computer Vision and Pattern Recognition, pages 248–255.
Elshennawy, N. M. and Ibrahim, D. M. (2020). Deep-pneumonia framework using deep learning models based on chest x-ray images. Diagnostics, 10(9).
Oliveira, P., de Carvalho Scabora, L., Cazzolato, M., Bedo, M., Traina, A., and Jr, C. (2017). Mammoset: An enhanced dataset of mammograms. In Dataset Showcase Workshop - DSW at the Brazilian Symposium on Databases, pages 1–11.
Orenstein, E., Beijbom, O., Peacock, E., and Sosik, H. (2015). Whoi-plankton-a large scale fine grained visual recognition benchmark dataset for plankton classification. Tech Report, pages 1–2.
Rosenberg, C., Hebert, M., and Schneiderman, H. (2005). Semi-supervised self-training of object detection models. In 2005 Seventh IEEE Workshops on Applications of Computer Vision - Volume 1, volume 1, pages 29–36.
T.R., S., Thomas, S. A., Kamath, V., and B., N. K. (2019). Hybrid xception model for human protein atlas image classification. In IEEE 16th India Council International Conference, pages 1–4.
van Engelen, J. E. and Hoos, H. H. (2019). A survey on semi-supervised learning. Machine Learning, 109:373–440.
Zhang, L., Lu, L., Nogues, I., Summers, R. M., Liu, S., and Yao, J. (2017). Deeppap: Deep convolutional networks for cervical cell classification. IEEE Journal of Biomedical and Health Informatics, 21:1633–1643.
Zhuang, F., Qi, Z., Duan, K., Xi, D., Zhu, Y., Zhu, H., Xiong, H., and He, Q. (2019). A comprehensive survey on transfer learning. CoRR, abs/1911.02685.
Publicado
31/07/2022
Como Citar
BASSANI, Cristiano N. de O.; SAITO, Priscila T. M.; BUGATTI, Pedro H..
Avaliação de abordagens semi-supervisionadas aplicadas a redes neurais convolucionais. In: BRAZILIAN E-SCIENCE WORKSHOP (BRESCI), 16. , 2022, Niterói.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2022
.
p. 25-32.
ISSN 2763-8774.
DOI: https://doi.org/10.5753/bresci.2022.222746.
