Frame and Feature-Context Video Super-Resolution

Authors: Bo Yan, Chuming Lin, Weimin Tan5597-5604

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive evaluations and comparisons demonstrate that our approach produces state-of-the-art results on a standard benchmark dataset, with advantages in terms of accuracy, efficiency, and visual quality over the existing approaches.
Researcher Affiliation Academia Bo Yan, Chuming Lin, Weimin Tan School of Computer Science, Shanghai Key Laboratory of Intelligent Information Processing, Fudan University {byan, cmlin17, wmtan14}@fudan.edu.cn
Pseudocode Yes This processing flow is summarized in Algorithms 1. Algorithm 1 Frame and Feature-Context Video Super Resolution
Open Source Code No The paper does not provide concrete access to its source code, nor does it explicitly state that its code is open-source or available.
Open Datasets Yes Our training dataset consists of 2 high-resolution videos (4k, 60fps): Venice and Myanmar downloaded from harmonic1. [footnote 1: https://www.harmonicinc.com/free-4k-demo-footage/] and standard Vid4 benchmark dataset (Liu and Sun 2011)
Dataset Splits No The paper mentions a training dataset and a benchmark dataset used for evaluation but does not specify explicit training, validation, and test splits with percentages or counts.
Hardware Specification Yes All experiments are carried out for 4x upscaling. We conduct our experiments on a machine with an Intel i7-7700k CPU and an Nvidia GTX 1080Ti GPU.
Software Dependencies No Our framework is implemented on the Tensor Flow platform. (No version number specified for TensorFlow or any other software dependency).
Experiment Setup Yes The parameters are updated with initial learning rate of 10 4 before 300K iteration steps and changed to 10 5 at the following 50K. The loss is minimized using Adam optimizer (Kingma and Ba 2015) and back-propagated through both networks NETL and NETC as well as through time.