Produção Nacional
Revisado por pares
Semantic segmentation with labeling uncertainty and class imbalance applied to vegetation mapping
2022; Elsevier BV; Volume: 108; Linguagem: Inglês
10.1016/j.jag.2022.102690
ISSN1872-826X
AutoresPatrik Olã Bressan, José Marcato, José Augusto Correa Martins, Maximilian Jaderson de Melo, Diogo Nunes Gonçalves, Daniel Matte Freitas, Ana Paula Marques Ramos, Michelle Taís Garcia Furuya, Lucas Prado Osco, Jonathan de Andrade Silva, Zhipeng Luo, Raymundo Cordero, Lingfei Ma, Jonathan Li, Wesley Nunes Gonçalves,
Tópico(s)Species Distribution and Climate Change
ResumoRecently, methods based on Convolutional Neural Networks (CNN) achieved impressive success in semantic segmentation tasks. However, challenges such as the class imbalance and the uncertainty in the pixel-labeling process are not completely addressed. As such, we present a new approach that calculates a weight for each pixel considering its class and uncertainty during the labeling process. The pixel-wise weights are used during training to increase or decrease the importance of the pixels. Experimental results show that the proposed approach leads to significant improvements in three challenging segmentation tasks in comparison to baseline methods. It was also proved to be more invariant to noise. The approach presented here may be used within a wide range of semantic segmentation methods to improve their robustness.
Referência(s)