Topology-Aware Segmentation Using Discrete Morse Theory
Authors: Xiaoling Hu, Yusu Wang, Li Fuxin, Dimitris Samaras, Chao Chen
ICLR 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | On diverse datasets, our method achieves superior performance on both the DICE score and topological metrics. ... Our method outperforms state-of-the-art methods in multiple topology-relevant metrics (e.g., ARI and VOI) on various 2D and 3D benchmarks. |
| Researcher Affiliation | Academia | Xiaoling Hu Stony Brook University Yusu Wang University of California, San Diego Li Fuxin Oregon State University Dimitris Samaras Stony Brook University Chao Chen Stony Brook University |
| Pseudocode | No | The paper does not contain a clearly labeled 'Pseudocode' or 'Algorithm' block. The methodology is described in narrative text with mathematical formulations. |
| Open Source Code | No | The paper does not provide a direct link to open-source code for the methodology nor does it explicitly state that the code is being released or made available. |
| Open Datasets | Yes | Six natural and biomedical 2D datasets are used: ISBI12 (Arganda Carreras et al., 2015), ISBI13 (Arganda-Carreras et al., 2013), CREMI, Crack Tree (Zou et al., 2012), Road (Mnih, 2013) and DRIVE (Staal et al., 2004). ... We use three different biomedical 3D datasets: ISBI13, CREMI and 3Dircadb (Soler et al., 2010). |
| Dataset Splits | Yes | For all the experiments, we use a 3-fold cross-validation to tune hyperparameters for both the proposed method and other baselines, and report the mean performance over the validation set. |
| Hardware Specification | No | The paper does not explicitly describe the specific hardware (e.g., GPU models, CPU types) used to run the experiments. It only mentions using U-Nets. |
| Software Dependencies | No | The paper mentions using a '2D U-net' and '3D U-Net' architectures, but it does not specify any software dependencies with version numbers (e.g., PyTorch 1.9, TensorFlow 2.x, Python 3.x). |
| Experiment Setup | Yes | Our loss has two terms, the cross-entropy term, Lbce and the DMT-loss, Ldmt: L(f, g) = Lbce(f, g)+ βLdmt(f, g), in which f is the likelihood, g is the ground truth, and β is the weight of Ldmt. ... In practice, we first pretrain the network with only the cross-entropy loss, and then train the network with the combined loss. ... For all the experiments, we use a 3-fold cross-validation to tune hyperparameters for both the proposed method and other baselines... When β = 3, the proposed DMT-loss achieves best performance 0.982 (Betti Error). |