Search Results for author:
Found 3 papers, 3 papers with code
DAEMA: Denoising Autoencoder with Mask Attention
Missing data is a recurrent and challenging problem, especially when using machine learning algorithms for real-world applications.
Anomaly Detection: How to Artificially Increase your F1-Score with a Biased Evaluation Protocol
Moreover, we show that the F1-score and the AVPR cannot be used to compare performances on different datasets as they do not reflect the intrinsic difficulty of modeling such data.
Residual Conv-Deconv Grid Network for Semantic Segmentation
However, for semantic image segmentation, where the task consists in providing a semantic class to each pixel of an image, feature maps reduction is harmful because it leads to a resolution loss in the output prediction.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
