IDPT: Interconnected Dual Pyramid Transformer for Face Super-Resolution
Authors: Jingang Shi, Yusi Wang, Songlin Dong, Xiaopeng Hong, Zitong Yu, Fei Wang, Changxin Wang, Yihong Gong
IJCAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments and visualizations on various datasets demonstrate the superiority of the proposed method for face super-resolution tasks. 4 Experiments |
| Researcher Affiliation | Academia | Jingang Shi1 , Yusi Wang1 , Songlin Dong1 , Xiaopeng Hong1 , Zitong Yu2 , Fei Wang1 , Changxin Wang1 and Yihong Gong1 1School of Software Engineering, Xi an Jiaotong University 2School of Electrical and Electronic Engineering, Nanyang Technological University |
| Pseudocode | No | The paper describes the model architecture and training process in prose and diagrams but does not provide pseudocode or a formally labeled algorithm block. |
| Open Source Code | No | We also thank Mindspore for the support of this work.2 https://www.mindspore.cn/ (This link is to the Mindspore framework, not the specific code for the paper's methodology, and the text does not state that their code is open-source or available at this link.) |
| Open Datasets | Yes | The Celeb A dataset [Liu et al., 2015] is utilized to train IDPT in the experiments. ... and 50 images from Helen dataset [Le et al., 2012] to conduct the experiments. |
| Dataset Splits | No | The paper describes how the training and testing sets are constructed and selected, but it does not specify explicit percentages or sample counts for training/validation/test splits, nor does it explicitly mention a dedicated validation set split. |
| Hardware Specification | No | The paper does not provide specific hardware details such as GPU models, CPU types, or memory amounts used for running experiments. |
| Software Dependencies | No | The experiments are implemented on Mindspore and Pytorch. (No version numbers provided for software dependencies, which is required for reproducibility.) |
| Experiment Setup | Yes | For UPFE, it produces feature maps with channel number of 32 for the first encoder stage, while the parameter α of Leaky ReLU is set to 0.01. In the dual pyramid structure, the maximum number of stages L is set to 4. ... For STMs, the attention head number and window size are set to 2i − 1 (i = 1, ..., L) and 8 × 8, respectively. We train IDPT by AdamW optimizer with β1 = 0.9, β2 = 0.99, ϵ = 10−8, and weight decay is set to 0.02. The learning rate is 2 × 10−4 and the batchsize is set to 16. Loss Setting. For IDPT, we set λAdv = 0 and λPcp = 0. For IDPT-GAN, we empirically set λAdv = 0.005 and λPcp = 0.1. |