Fully Convolutional Network for Consistent Voxel-Wise Correspondence

Authors: Yungeng Zhang, Yuru Pei, Yuke Guo, Gengyu Ma, Tianmin Xu, Hongbin Zha12935-12942

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on both synthetic and clinically captured volumetric cone-beam CT (CBCT) images show that the proposed framework is effective and competitive against state-of-the-art deformable registration techniques.
Researcher Affiliation Collaboration 1Key Laboratory of Machine Perception (MOE), Department of Machine Intelligence, Peking University, Beijing, China 2 Luoyang Institute of Science and Technology, Luoyang, China 3 u Sens Inc., San Jose, USA 4 School of Stomatology, Peking University, Beijing, China
Pseudocode No The paper describes the proposed method in text and diagrams, but does not include any explicit pseudocode or algorithm blocks.
Open Source Code No The paper does not provide an explicit statement about releasing open-source code or a link to a code repository for their method.
Open Datasets No The paper states: 'The training dataset consists of 400 clinically captured CBCT images from orthodontic patients...' and 'we generate a toy dataset with the ground-truth DVFs using synthetic data...'. However, no specific link, DOI, or formal citation to a publicly available version of these datasets is provided.
Dataset Splits No The paper states, 'The training dataset consists of 400 clinically captured CBCT images...' and 'For testing, we collect a toy dataset with 20 synthetic images and a real dataset with 20 clinically captured images.' It does not explicitly define a separate validation dataset split.
Hardware Specification Yes The framework is implemented using the open-source Py Torch implementation of convolutional neural networks on an NVIDIA GTX TITAN X GPU.
Software Dependencies No The paper mentions 'Py Torch implementation' but does not specify a version number for PyTorch or any other software dependencies.
Experiment Setup Yes We train the network using the ADAM optimizer with a learning rate of 1e-4 and momentums of 0.5 and 0.999. The mini-batch contains three volumes.