Research map / Learning theory and algorithms / Matrix and tensor methods
Solving Phase Retrieval with a Learned Reference
Rakib Hyder, Zikui Cai, M. Salman Asif
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add a learned reference rather than none (HIO) | Phase retrieval · MNIST | PSNR (dB): 9.04 -> 66.54 | Improved |
| learn the reference instead of a random reference (ablation) | Phase retrieval · CIFAR10 | PSNR (dB): 25.57 -> 41.60 | Improved |
| learn the reference instead of a random reference (ablation) | Phase retrieval · MNIST | PSNR (dB): 36.87 -> 66.54 | Improved |
| learn the reference instead of a flat reference (ablation) | Phase retrieval · MNIST | PSNR (dB): 18.21 -> 66.54 | Improved |
| increase training size from 32 to 128 and 512 images (ablation) | Learned Reference · MNIST | PSNR (dB): 66.54 -> 76.25 -> 79.14 | Improved |
| increase training size from 32 to 512 images (ablation) | Learned Reference · Fashion MNIST | PSNR (dB): 57.81 -> 52.01 | Hurt |
| increase training size from 32 to 512 images (ablation) | Learned Reference · CIFAR10 | PSNR (dB): 41.60 -> 48.90 | Improved |
| train reference on a different dataset than the test one (ablation) | Learned Reference · CIFAR10 test | PSNR (dB): 41.60 -> 31.72 (MNIST reference) | Mixed |
| train reference on SVHN rather than CIFAR10 itself (ablation) | Learned Reference · CIFAR10 test | PSNR (dB): 41.60 -> 51.54 | Improved |
| increase number of unrolled layers K (ablation) | Learned Reference · several datasets | PSNR (dB) | Improved |
| place 8x8 reference at image center instead of corner (ablation) | Learned Reference · MNIST/CIFAR10 | PSNR (dB) | Hurt |
| use larger value range for the reference (ablation) | Learned Reference · several datasets | PSNR (dB) | Improved |
Closest papers by meaning
- DeepPhaseCut: Deep Relaxation in Phase for Unsupervised Fourier Phase Retrieval (TPAMI 2021 · similarity 0.89)
- [Re] Solving Phase Retrieval With a Learned Reference (NeurIPS 2022 · similarity 0.89)
- When deep denoising meets iterative phase retrieval (ICML 2020 · similarity 0.88)
- DiffFPR: Diffusion Prior for Oversampled Fourier Phase Retrieval (ICML 2024 · similarity 0.86)
- A Fast and Provable Algorithm for Sparse Phase Retrieval (ICLR 2024 · similarity 0.86)
- Exponential Spectral Pursuit: An Effective Initialization Method for Sparse Phase Retrieval (ICML 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
