Real-World Deep Local Motion Deblurring

Authors: Haoying Li, Ziran Zhang, Tingting Jiang, Peng Luo, Huajun Feng, Zhihai Xu

AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments prove the reliability of Re Lo Blur dataset, and demonstrate that LBAG achieves better performance than state-of-the-art global deblurring methods and our proposed local blur-aware techniques are effective.
Researcher Affiliation Academia 1 College of Optical Science and Engineering, Zhejiang University 2 Research Center for Intelligent Sensing Systems, Zhejiang Laboratory {lhaoying, naturezhanghn, luop, fenghj, xuzh}@zju.edu.cn, eagerjtt@zhejianglab.com
Pseudocode No The paper describes methods with figures and textual explanations but does not include structured pseudocode or algorithm blocks.
Open Source Code No The paper states "Re Lo Blur contains 2405 image pairs and we will release the dataset soon" and provides a project homepage link (https://leiali.github.io/Re Lo Blur homepage/index.html), but there is no explicit statement or link confirming the release of the source code for the described methodology.
Open Datasets Yes We establish the first real local motion blur dataset, Re Lo Blur, captured by a synchronized beam-splitting photographing system in daily real scenes, and corrected by a post-processing pipeline. Re Lo Blur contains 2405 image pairs and we will release the dataset soon. https://leiali.github.io/Re Lo Blur homepage/index.html
Dataset Splits No The paper states "We split the Re Lo Blur dataset into 2010 pairs for training and 395 pairs for testing" but does not explicitly provide details for a validation split.
Hardware Specification Yes For a fair comparison, we trained LBAG and the baseline deblurring methods for the same steps on 1 Ge Force RTX 3090 with 24GB of memory.
Software Dependencies No The paper mentions specific optimizers and methods (e.g., Adam, Pyflow, CMTF) but does not provide version numbers for any software dependencies or libraries.
Experiment Setup Yes We crop the images to 256 × 256 patches as the training inputs using BAPC strategy. For data augmentation, each patch is horizontally or vertically flipped with a probability of 0.5. We use Adam (Kingma and Ba 2014) as the optimizer, with a batchsize of 12 and an initial learning rate of 10−4, which is halved every 100k steps. The training procedure takes approximately 70 hours (300k steps).