High-Order Residual Network for Light Field Super-Resolution

Authors: Nan Meng, Xiaofei Wu, Jianzhuang Liu, Edmund Lam11757-11764

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results show that our approach enables high-quality reconstruction even in challenging regions and outperforms state-of-the-art single image or LF reconstruction methods with both quantitative measurements and visual evaluation.
Researcher Affiliation Collaboration Nan Meng,1 Xiaofei Wu,2 Jianzhuang Liu,2 Edmund Y. Lam1 1Dept. of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong 2Huawei Noah s Ark Lab, China
Pseudocode Yes Algorithm 1: Aperture group batch normalization
Open Source Code Yes The entire implementation is available at https://github.com/monaen/Light Field Reconstruction.
Open Datasets Yes In the experiments, we randomly select 100 scenes from the Lytro Archive (Stanford) (excluding Occlusions and Reflective ) and the entire Fraunhofer densely-sampled highresolution (Ziegler et al. 2017) datasets for training.
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, or detailed methodology) for validation data.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes During training, the system each time receives a 4D patch of LF, which is spatially cropped to 96 x 96 as input. For spatial SR, the downsampling is based on the classical model (Farrugia and Guillemot 2018) ... The network is trained using the Stochastic Gradient Descent solver with the initial learning rate of 10 5, which is decreased by a factor of 0.1 for every 10 epochs.