Learning to reconstruct from saturated data: audio declipping and high-dynamic range imaging
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917294867742720 |
|---|---|
| author | Sechaud, Victor Jacques, Laurent Abry, Patrice Tachella, Julián |
| author_facet | Sechaud, Victor Jacques, Laurent Abry, Patrice Tachella, Julián |
| contents | Learning based methods are now ubiquitous for solving inverse problems, but their deployment in real-world applications is often hindered by the lack of ground truth references for training. Recent self-supervised learning strategies offer a promising alternative, avoiding the need for ground truth. However, most existing methods are limited to linear inverse problems. This work extends self-supervised learning to the non-linear problem of recovering audio and images from clipped measurements, by assuming that the signal distribution is approximately invariant to changes in amplitude. We provide sufficient conditions for learning to reconstruct from saturated signals alone and a self-supervised loss that can be used to train reconstruction networks. Experiments on both audio and image data show that the proposed approach is almost as effective as fully supervised approaches, despite relying solely on clipped measurements for training. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_22279 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Learning to reconstruct from saturated data: audio declipping and high-dynamic range imaging Sechaud, Victor Jacques, Laurent Abry, Patrice Tachella, Julián Image and Video Processing Artificial Intelligence Sound Learning based methods are now ubiquitous for solving inverse problems, but their deployment in real-world applications is often hindered by the lack of ground truth references for training. Recent self-supervised learning strategies offer a promising alternative, avoiding the need for ground truth. However, most existing methods are limited to linear inverse problems. This work extends self-supervised learning to the non-linear problem of recovering audio and images from clipped measurements, by assuming that the signal distribution is approximately invariant to changes in amplitude. We provide sufficient conditions for learning to reconstruct from saturated signals alone and a self-supervised loss that can be used to train reconstruction networks. Experiments on both audio and image data show that the proposed approach is almost as effective as fully supervised approaches, despite relying solely on clipped measurements for training. |
| title | Learning to reconstruct from saturated data: audio declipping and high-dynamic range imaging |
| topic | Image and Video Processing Artificial Intelligence Sound |
| url | https://arxiv.org/abs/2602.22279 |