Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
FIPER: Factorized Features for Robust Image Super-Resolution and Compression
Authors: Yang-Che Sun, Cheng-Yu Yeo, Ernie Chu, Jun-Cheng Chen, Yu-Lun Liu
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments show that our unified representation delivers state-of-the-art performance, achieving an average relative improvement of 204.4% in PSNR over the baseline in Super-Resolution (SR) and 9.35% BD-rate reduction in Image Compression compared to the previous SOTA. Project page: https://jayisaking.github.io/FIPER/ |
| Researcher Affiliation | Academia | Yang-Che Sun1 Cheng Yu Yeo1 Ernie Chu2 Jun-Cheng Chen3 Yu-Lun Liu1 1National Yang Ming Chiao Tung University 2Johns Hopkins University 3Academia Sinica |
| Pseudocode | No | The paper describes methods and formulations but does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | Question: Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material? Answer: [Yes] Justification: We will make a full open access to both the code and the data used in our experiments. |
| Open Datasets | Yes | table 1 presents the quantitative comparison between our approach and state-of-the-art (So TA) methods. We evaluate the methods using five benchmark datasets, including Set5 [6], Set14 [107], BSD100 [75], Urban100 [38], and Manga109 [76]. |
| Dataset Splits | Yes | The training dataset includes the REDS[82] and Vimeo90K[103] datasets, while the testing dataset comprices REDS4[82], Vid4[60], and Vimeo-90KT[103]. |
| Hardware Specification | Yes | Question: For each experiment, does the paper provide sufficient information on the computer resources (type of compute workers, memory, time of execution) needed to reproduce the experiments? Answer: [Yes] Justification: We use GPU GTX 4090 as our computational resources. |
| Software Dependencies | No | The paper mentions optimizers like Adam W and specific versions of datasets but does not specify software dependencies like programming languages, libraries, or frameworks with their version numbers. |
| Experiment Setup | Yes | Pretraining runs 300k iterations on Image Net with Adam W (lr 2e-4, batch 32, β = 0.9/0.99); finetuning runs 200k iterations on DF2K (DIV2K [1]+Flickr2K [59]) with lr 1e-5. Inputs are 256 256 crops, bicubically down-sampled to 64 64 for the backbone. We set the number of coefficient basis pairs to N = 6 and use frequency scalars ιj {1, 4, 16, 64} to capture both lowand high-frequency details. |