Decoupling with Entropy-based Equalization for Semi-Supervised Semantic Segmentation
Authors: Chuanghao Ding, Jianrong Zhang, Henghui Ding, Hongwei Zhao, Zhihui Wang, Tengfei Xing, Runbo Hu
IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct extensive experiments of the proposed De S4 on two challenging benchmarks (PASCAL VOC 2012 and Cityscapes) and achieve remarkable improvements over the previous state-of-the-art methods. |
| Researcher Affiliation | Collaboration | Chuanghao Ding1,2,5, , Jianrong Zhang1,3 , Henghui Ding4 , Hongwei Zhao1,3, , Zhihui Wang5 , Tengfei Xing5, and Runbo Hu5 1Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, China 2 College of Software, Jilin University, China 3 College of Computer Science and Technology, Jilin University, China 4Nanyang Technological University, Singapore 5Didi Chuxing, China |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide a statement or link indicating access to the source code for the described methodology. |
| Open Datasets | Yes | PASCAL VOC 2012 [Everingham et al., 2015] is the most widely used benchmark dataset in semi-supervised semantic segmentation... Cityscapes [Cordts et al., 2016] is a high-resolution urban scene dataset with a total of 19 classes. |
| Dataset Splits | Yes | PASCAL VOC 2012 [Everingham et al., 2015] is the most widely used benchmark dataset in semi-supervised semantic segmentation... The original dataset consists of 1464 images for training and 1449 images for evaluation... We follow previous works [Zhang et al., 2022] to select 1/4 and 1/8 training images as labeled data... All the experimental results are evaluated on either the VOC Val set or the Cityscapes Val set, and ablation studies are conducted on the 1/4 and 1/8 VOC Aug dataset. |
| Hardware Specification | Yes | All experiments are trained on 8 NVIDIA RTX A6000 GPUs with a batch size of 16 |
| Software Dependencies | No | The paper mentions using 'Deep Lab v3+ [Chen et al., 2018b] as the semantic segmentation network with the Res Net101 backbone' but does not specify software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow, CUDA versions). |
| Experiment Setup | Yes | We use Deep Lab v3+ [Chen et al., 2018b] as the semantic segmentation network with the Res Net101 backbone. All experiments are trained on 8 NVIDIA RTX A6000 GPUs with a batch size of 16, and we use stochastic gradient descent (SGD) to optimize the model, and set balance weights λ1 and λ2 to 1 and 1.5. Empirically, we set both the EMA decay of τ1 and τ2 to 0.99. For the Multi-Entropy Sampling strategy, we set k and α to 5 and 0.1, respectively. For both PASCAL VOC Train and Aug datasets, the initial learning rate is set to 0.001, and the weight decay is 0.0001. We follow previous settings [Wang et al., 2022] to train our model for 80 epochs with the crop size of 513 513. For the Cityscapes dataset, the initial learning rate is 0.01, weight decay is 0.0006, and the crop size is 769 769. Furthermore, we employ a poly learning rate policy that the initial learning rate is multiplied by (1 iter max iter)power with power = 0.9. |