Double-Check Soft Teacher for Semi-Supervised Object Detection
Authors: Kuo Wang, Yuxiang Nie, Chaowei Fang, Chengzhi Han, Xuewen Wu, Xiaohui Wang Wang, Liang Lin, Fan Zhou, Guanbin Li
IJCAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Double Check Soft Teacher consistently surpasses state-of-the-art methods by significant margins on the MS-COCO benchmark, pushing the new state-of-the-art. |
| Researcher Affiliation | Collaboration | 1School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China 2School of Artificial Intelligence, Xidian University, Xi an, China 3Huawei Technologies Co. China |
| Pseudocode | No | The paper describes the algorithm steps in text and provides a system diagram (Figure 2) but does not include structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Source codes are available at https://github.com/wkfdb/DCST. |
| Open Datasets | Yes | Following existing works, we test the performance of our method on MS-COCO benchmark [Lin et al., 2014]. Subsets train2017 and unlabeled2017 are used for training |
| Dataset Splits | Yes | In partially labeled data protocol, we respectively sample 1%, 5% and 10% of train2017 as labeled data, and the remaining is used as unlabeled data. For every setting, five trials are conducted using different random seeds for random data sampling, and we report the averaged evaluation metrics. |
| Hardware Specification | No | For partially labeled data, the model is trained for 180,000 iterations on 4 GPUs. Each GPU processes ten images. ... For fully labeled data, the model is trained for 720,000 iterations on 8 GPUs. Each GPU processes eight images. The paper mentions the number of GPUs but does not specify their model or type. |
| Software Dependencies | No | The code is implemented under the framework of MMdetection [Chen et al., 2019]. The paper mentions MMdetection but does not specify its version number or other software dependencies with versions. |
| Experiment Setup | Yes | For partially labeled data, the model is trained for 180,000 iterations on 4 GPUs. Each GPU processes ten images. The learning rate is initialized to 0.1 and is divided by 10 at the 120,000-th and 160,000-th iteration. ... For fully labeled data, the model is trained for 720,000 iterations on 8 GPUs. Each GPU processes eight images. λ is set to 2. The learning rate is initialized as 0.1 and is divided by 10 at the 480,000-th and 640,000-th iteration. |