Shape-Aware Organ Segmentation by Predicting Signed Distance Maps
Authors: Yuan Xue, Hui Tang, Zhi Qiao, Guanzhong Gong, Yong Yin, Zhen Qian, Chao Huang, Wei Fan, Xiaolei Huang12565-12572
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We validate our proposed models by conducting extensive experiments on a hippocampus segmentation dataset and the public MICCAI 2015 Head and Neck Auto Segmentation Challenge dataset with multiple organs. We report the quantitative results of 3 models in Table 1. To validate our proposed methods, we conduct comprehensive experiments on a single organ segmentation dataset and a multi-organ segmentation dataset. |
| Researcher Affiliation | Collaboration | Yuan Xue,1 Hui Tang,2 Zhi Qiao,2 Guanzhong Gong,3 Yong Yin,3 Zhen Qian,2 Chao Huang,2 Wei Fan,2 Xiaolei Huang1 1Pennsylvania State University, University Park, PA, USA 2Tencent Hippocrates Research Lab, Palo Alto, CA, USA 3Shandong Cancer Hospital and Institute, Jinan, Shandong, China |
| Pseudocode | No | The paper does not contain any pseudocode or clearly labeled algorithm blocks. |
| Open Source Code | No | The paper does not provide any statement or link indicating the availability of open-source code for the methodology described. |
| Open Datasets | Yes | To validate our proposed methods, we conduct comprehensive experiments on a single organ segmentation dataset and a multi-organ segmentation dataset. The single organ dataset is our collected hippocampus segmentation dataset. ... To examine the effectiveness of our proposed method on the more challenging multi-organ segmentation task and compare with current state-of-the-art organ segmentation algorithms, we further conduct experiments on the MICCAI Head and Neck Auto Segmentation Challenge 2015 dataset (Raudaschl et al. 2017). |
| Dataset Splits | No | The paper specifies training and testing splits for both datasets (60 training, 12 testing for hippocampus; 38 training, 10 testing for MICCAI 2015), but does not explicitly detail a separate validation split or how validation was handled for hyperparameter tuning or early stopping. |
| Hardware Specification | Yes | All experiments are done on a single NVIDIA Tesla P40 GPU with 24G memory. |
| Software Dependencies | No | The paper mentions optimizers, activation functions, and normalization types, but does not provide specific version numbers for any software libraries or frameworks (e.g., Python, TensorFlow, PyTorch, CUDA versions). |
| Experiment Setup | Yes | In our proposed deep 3D UNet, the initial number of channels is 24 and is doubled in each downsampling operation. The maximum number of channels is 384. All models are trained by Adam optimizer. The initial learning rate is 5e 4 and decayed by factor of 0.8 for every 25 epochs. All models are trained for 200 epochs for the Hippocampus dataset and 600 epochs for the MICCAI 2015 dataset. The batchsize is 1 and all experiments are done on a single NVIDIA Tesla P40 GPU with 24G memory. During inference, the segmentation result is obtained by applying the Heaviside function to the predicted SDM. where the final value of λ is set to 10 in all experiments. |