Controllable Image-to-Video Translation: A Case Study on Facial Expression Generation

Authors: Lijie Fan, Wenbing Huang, Chuang Gan, Junzhou Huang, Boqing Gong3510-3517

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments and user studies verify the effectiveness of our approach.
Researcher Affiliation Collaboration Lijie Fan,1 Massachusetts Institute of Technology, 2Tencent AI Lab, 3MIT-Watson Lab lijiefan@mit.edu, hwenbing@126.com, chuangg@mit.edu, jzhuang@uta.edu, boqinggo@outlook.com
Pseudocode No The paper does not contain a block labeled 'Pseudocode' or 'Algorithm'.
Open Source Code No The paper does not provide any specific statement or link indicating that the source code for their methodology is publicly available.
Open Datasets Yes We not only use the public CK+ (Lucey et al. 2010) dataset for model training but also significantly extend it in scale. The new larger-scale dataset is named CK++.
Dataset Splits Yes We use 10 video clips from the CK++ dataset for validation and all the others for training.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., CPU, GPU models, or memory) used for running the experiments.
Software Dependencies No The paper mentions using 'Dlib Library (King 2009)' but does not specify its version number or versions for any other software dependencies.
Experiment Setup Yes The Adam optimizer is used in the experiments, with the initial learning rate of 0.0002. The whole training process takes 2100 epochs, where one epoch means a complete pass over the training data. All images are resized to 289x289 and randomly cropped to 256x256 before being fed into the network. We set the small increment to a = 0.1 for temporal regulation Rt.