Light-Weight Hybrid Convolutional Network for Liver Tumor Segmentation
Authors: Jianpeng Zhang, Yutong Xie, Pingping Zhang, Hao Chen, Yong Xia, Chunhua Shen
IJCAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluated the proposed LW-HCN model against several recent methods on the Li TS and 3D-IRCADb datasets and achieved, respectively, the Dice per case of 73.0% and 94.1% for tumor segmentation, setting a new state of the art. |
| Researcher Affiliation | Academia | 1School of Computer Science and Engineering, Northwestern Polytechnical University, P.R. China 2School of Information and Communication Engineering, Dalian University of Technology, P.R. China 3School of Computer Science, University of Adelaide, Adelaide, SA 5005, Australia |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide concrete access to source code (specific repository link, explicit code release statement, or code in supplementary materials) for the methodology described. |
| Open Datasets | Yes | We evaluated the LW-HCN model on the Li TS dataset and 3D-IRCADb dataset. The Li TS dataset is composed of 201 contrast-enhanced abdominal CT volumes... Following the evaluation procedures of the Li TS challenge1, we evaluated the segmentation performance... (footnote: https://competitions.codalab.org/competitions/17094#learn the details) |
| Dataset Splits | Yes | The Li TS dataset is composed of 201 contrast-enhanced abdominal CT volumes provided by various clinical sites around the world, including 131 volumes for training and 70 volumes for testing. ... We select 5 volumes from the Li TS training data to form a validation set, which is used to monitor the performance of our model. |
| Hardware Specification | Yes | Our model is implemented with Keras and optimized with the Adam algorithm [Kingma and Ba, 2015] on a NVIDIA Tesla P100 GPU. |
| Software Dependencies | No | The paper mentions 'Keras' as the implementation framework and 'Adam algorithm' for optimization, but does not provide specific version numbers for these or other software dependencies. |
| Experiment Setup | Yes | The initial learning rate is set to 0.001 and decayed according to the poly schedule lr = lr (1 iterations/total iterations)0.9. ... During training, we densely sample 12 256 256 sub-volumes from each CT scan as the input of the model. ... we adopt the gradient-checkpointing algorithm [Chen et al., 2016] to enlarge the batch size to 6. |