Generative Adversarial Networks for Video-to-Video Domain Adaptation

Authors: Jiawei Chen, Yuexiang Li, Kai Ma, Yefeng Zheng3462-3469

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results demonstrate that the adapted colonoscopic video generated by our Video GAN can significantly boost the segmentation accuracy, i.e., an improvement of 5%, of colorectal polyps on multicentre datasets.
Researcher Affiliation Industry Jiawei Chen, Yuexiang Li, Kai Ma, Yefeng Zheng You Tu Lab, Tencent, Shenzhen, China {jiaweichen, vicyxli, kylekma, yefengzheng}@tencent.com
Pseudocode No The paper provides network diagrams but no structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide an explicit statement about the release of its source code or a link to a code repository.
Open Datasets Yes The publicly available colonoscopic video datasets, i.e., CVC-Clinic4 (V azquez et al. 2017) and ETISLarib5 (Silva et al. 2014), are selected for our experiments.
Dataset Splits No The paper does not explicitly provide specific dataset split information for training, validation, or testing, nor does it specify the methodology for such splits.
Hardware Specification Yes The network is trained with a minibatch size of 1 on one GPU (Tesla P40 with 24 GB memory).
Software Dependencies No The proposed Video GAN is implemented using Py Torch. The paper does not provide specific version numbers for PyTorch or other software dependencies.
Experiment Setup Yes The Adam solver (Kingma and Ba 2014) with betas = (0.5, 0.999) is adopted for the optimization of Video GAN. The network is trained with a minibatch size of 1 on one GPU (Tesla P40 with 24 GB memory). The initial learning rate is set to 0.0002. The proposed Video GAN yields visually satisfactory translated frames after 200 training epochs.