Towards Scale-Free Rain Streak Removal via Self-Supervised Fractal Band Learning
Authors: Wenhan Yang, Shiqi Wang, Dejia Xu, Xiaodong Wang, Jiaying Liu12629-12636
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments in quantitative and qualitative evaluations demonstrate the superiority of our FBL for rain streak removal, especially for the real cases where very large rain streaks exist, and prove the effectiveness of its each component. |
| Researcher Affiliation | Collaboration | 1City University of Hong Kong 2Peking University 3Beijing Institute of Electronic Engineering |
| Pseudocode | No | The paper does not contain any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Our code will be public available at: https://github.com/flyywh/AAAI-2020-FBL-SS. |
| Open Datasets | Yes | We compare our method with state-of-the-art methods on a few benchmark datasets: (1) Rain100L and Rain100H (Yang et al. 2017b), which are synthesized datasets with only one type of rain streaks and with three to five layers of rain streaks, respectively; (2) Rain100H-S2 and Rain100-S3 proposed in (Yang et al. 2019), synthesized with s rain streaks (s {2, 3, 4, 5}) with different shapes and directions. The streak sizes are twice and three times as large as those in Rain100H, used for evaluating the performance when training and testing streaks have different sizes. (3) Rain800, a collection of diversified synthesized rain images from randomly selected outdoor images, which is split into testing set of 100 image and training set of 700 images. |
| Dataset Splits | Yes | Rain800, a collection of diversified synthesized rain images from randomly selected outdoor images, which is split into testing set of 100 image and training set of 700 images. |
| Hardware Specification | No | The paper does not specify any particular hardware (e.g., GPU models, CPU types) used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific software names with version numbers for dependencies (e.g., Python 3.x, PyTorch 1.x). |
| Experiment Setup | No | The paper states 'For a fair comparison, we make sure the parameters of these methods are almost the same' but does not provide specific details on hyperparameters (e.g., learning rate, batch size, epochs) or other system-level training settings. |