Benchmarking the Fairness of Image Upsampling Methods

arXiv:2401.13555v2 Announce Type: replace-cross
Abstract: Recent years have witnessed a rapid development of deep generative models for creating synthetic media, such as images and videos. While the practical applications of these models in everyday tasks are enticing, it is crucial to assess the inherent risks regarding their fairness. In this work, we introduce a comprehensive framework for benchmarking the performance and fairness of conditional generative models. We develop a set of metrics$\unicode{x2013}$inspired by their supervised fairness counterparts$\unicode{x2013}$to evaluate the models …

abstract applications arxiv benchmarking cs.ai cs.cv cs.lg deep generative models development fairness generative generative models image images media practical risks synthetic synthetic media tasks type videos while work