SCMT: Self-Correction Mean Teacher for Semi-supervised Object Detection
Authors: Feng Xiong, Jiayi Tian, Zhihui Hao, Yulin He, Xiaofeng Ren
IJCAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on MS-COCO benchmark have shown the superiority of our proposed SCMT, which can significantly improve the supervised baseline by more than 11% m AP under all 1%, 5% and 10% COCO-standard settings, and surpasses state-of-the-art methods by about 1.5% m AP. Even under the challenging COCO-additional setting, SCMT still improves the supervised baseline by 4.9% m AP, and significantly outperforms previous methods by 1.2% m AP, achieving a new state-of-the-art performance. |
| Researcher Affiliation | Industry | Feng Xiong , Jiayi Tian , Zhihui Hao , Yulin He and Xiaofeng Ren Alibaba Group {xf250971, tianjiayi.tjy, hzh106945, harrylin95, x.ren}@alibaba-inc.com, |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper implements its framework based on an existing toolbox (MMDetection) but does not state that the code for their specific contribution (SCMT) is open-source or provide a link to it. |
| Open Datasets | Yes | We evaluate our approach on the large-scale dataset MSCOCO [Lin et al., 2014]. For MS-COCO, we use version 2017 in all experiments, including train2017, unlabeled2017 and val2017. |
| Dataset Splits | Yes | For MS-COCO, we use version 2017 in all experiments, including train2017, unlabeled2017 and val2017. train2017 set has a total of 118k labeled images, unlabeled2017 set contains 123k unlabeled images and val2017 set has 5k images in total. For COCO-standard setting, we set up three different proportions: 1%, 5%, and 10% to sample images from train2017 as the labeled training data, and the remaining unsampled images are used as the unlabeled data. For COCO-additional setting, we use the fully train2017 as the labeled data and the additional unlabeled2017 as the unlabeled data. Besides, val2017 set is used to evaluate our approach on both settings. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., CPU/GPU models, memory) used for running its experiments. It mentions "All models are trained with batch size of 32" but no hardware. |
| Software Dependencies | No | The paper states: "We implement our SCMT framework based on the MMDetection toolbox [Chen et al., 2019]." However, it does not specify the version number of MMDetection or any other software dependencies. |
| Experiment Setup | Yes | The filtering threshold is one of the main hyper-parameters in our SCMT framework, and is set to 0.7 throughout our experiments. For both settings, the initial learning rate is set to 0.01. For COCO-standard setting, the learning rate decays by 10 at 120k and 160k, with a total training step of 180k. For COCO-additional setting, the learning rate decays by 10 at 240k and 320k, with a total training step of 360k. All models are trained with batch size of 32, and the other training hyper-parameters are the same as standard Faster R-CNN. |