Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1]

Spatiotemporal-Aware Neural Fields for Dynamic CT Reconstruction

Authors: Qingyang Zhou, Yunfan Ye, Zhiping Cai

AAAI 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conducted experiments on medical and industrial datasets covering various motion types, sampling modes, and reconstruction resolutions. Experimental results show that our method outperforms the second-best by 5.99 d B and 4.11 d B in medical and industrial scenes, respectively.
Researcher Affiliation Academia 1College of Computer Science and Technology, National University of Defense Technology, Changsha, China 2School of Design, Hunan University, Changsha, China
Pseudocode No The paper describes the methodology using textual explanations and figures, but it does not contain a clearly labeled pseudocode or algorithm block.
Open Source Code Yes Code https://qingyangzhou69.github.io/STNF4D
Open Datasets Yes We collected 4D extended cardiactorso (XCAT) (Segars et al. 2008) phantom and real patient 4D-CT images from 4D-Lung Cancer Imaging Archive (TCIA) (Hugo et al. 2017), which were consistent with (Zhi et al. 2021). We collected the dataset (Reed et al. 2021) of aluminum products damage and evolution under various external forces, which provides valuable information for material performance and safety.
Dataset Splits No The paper describes the characteristics of the datasets and how projections were synthesized, but it does not explicitly provide training/test/validation dataset splits or refer to predefined splits in a way that would allow direct reproduction of data partitioning for model training and evaluation.
Hardware Specification Yes Our method is implemented based on Pytorch, and all experiments are completed on a single RTX 4090 GPU.
Software Dependencies No Our method is implemented based on Pytorch, and all experiments are completed on a single RTX 4090 GPU. While Pytorch is mentioned, a specific version number is not provided, nor are other software dependencies with their versions.
Experiment Setup Yes The model is optimized using the Adam optimizer. The initial learning rate is set to 1e-2, and the learning rate is reduced to 1e-6 using the cosine decay strategy. The number of iterations for each scene is 40k, and the number of rays per batch is 1024. In the medical dataset, the number of sampling points per ray is 768; in the industrial dataset, the number of sampling points per ray is 320.