@article{Alcover-Couso2026, 
author = {Roberto Alcover-Couso and Marcos Escudero-Viñolo and Juan C. SanMiguel and José M. Martinez},
title = {Soft-labelling for budget-constrained semantic segmentation: Bringing coherence to label down-sampling},
year = {2026},
journal = {Computational Visual Media},
volume = {12},
number = {3},
pages = {765-785},
keywords = {semantic segmentation, data augmentation, resource constraints, efficient training, data down-sampling},
url = {https://www.sciopen.com/article/10.26599/CVM.2025.9450470},
doi = {10.26599/CVM.2025.9450470},
abstract = {In semantic segmentation, training data down-sampling is commonly performed due to resource limitations, the need to adapt image size to the model input, or to improve data augmentation. This down-sampling typically employs different strategies for the image data and the annotated labels. Such discrepancy leads to mismatches between the down-sampled colour and ground-truth label images. Hence, the training performance significantly decreases as the down-sampling factor increases. In this paper, we bring together the down-sampling strategies for the image data and the training labels. To that aim, we propose a novel framework for label down-sampling via soft-labelling that better conserves label information after down-sampling, thereby, fully aligning soft-labels with image data to keep the distribution of the sampled pixels for down-sampling. This proposal also produces reliable annotations for under-represented semantic classes. Altogether, it allows training competitive models at lower resolutions. Experiments show that our proposal outperforms other down-sampling strategies. Moreover, state-of-the-art performance is achieved for reference benchmarks, but employing significantly fewer computational resources than foremost methods. This proposal enables competitive research for semantic segmentation under resource constraints.}
}