{"ID":2823115,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2601.01228","arxiv_id":"2601.01228","title":"HyDRA: Hybrid Denoising Regularization for Measurement-Only DEQ Training","abstract":"Solving image reconstruction problems of the form \\(\\mathbf{A} \\mathbf{x} = \\mathbf{y}\\) remains challenging due to ill-posedness and the lack of large-scale supervised datasets. Deep Equilibrium (DEQ) models have been used successfully but typically require supervised pairs \\((\\mathbf{x},\\mathbf{y})\\). In many practical settings, only measurements \\(\\mathbf{y}\\) are available. We introduce HyDRA (Hybrid Denoising Regularization Adaptation), a measurement-only framework for DEQ training that combines measurement consistency with an adaptive denoising regularization term, together with a data-driven early stopping criterion. Experiments on sparse-view CT demonstrate competitive reconstruction quality and fast inference.","short_abstract":"Solving image reconstruction problems of the form \\(\\mathbf{A} \\mathbf{x} = \\mathbf{y}\\) remains challenging due to ill-posedness and the lack of large-scale supervised datasets. Deep Equilibrium (DEQ) models have been used successfully but typically require supervised pairs \\((\\mathbf{x},\\mathbf{y})\\). In many practic...","url_abs":"https://arxiv.org/abs/2601.01228","url_pdf":"https://arxiv.org/pdf/2601.01228v1","authors":"[\"Markus Haltmeier\",\"Lukas Neumann\",\"Nadja Gruber\",\"Johannes Schwab\",\"Gyeongha Hwang\"]","published":"2026-01-03T16:28:05Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"math.NA\"]","methods":"[]","has_code":false}