iFAN: Image-Instance Full Alignment Networks for Adaptive Object Detection
Authors: Chenfan Zhuang, Xintong Han, Weilin Huang, Matthew Scott13122-13129
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate i FAN on two domain adaptation scenarios: 1) train on SIM10K (Johnson-Roberson et al. 2017) and test on Cityscapes (Cordts et al. 2016) dataset (SIM10K Cityscapes); 2) train on Cityscapes (Cordts et al. 2016) and test on Foggy Cityscapes (Sakaridis, Dai, and Van Gool 2018) (Cityscapes Foggy). ... Our method is compared with state-of-the-art UDA object detectors in Table 1. ... Ablation Study. We conduct ablation study by isolating each component in i FAN. The results are presented in Table 2 and 3. |
| Researcher Affiliation | Industry | Chenfan Zhuang, Xintong Han, Weilin Huang, Matthew R. Scott Malong Technologies, Shenzhen, China Shenzhen Malong Artificial Intelligence Research Center, Shenzhen, China {fan, xinhan, whuang, mscott}@malong.com |
| Pseudocode | No | The paper does not contain any explicit pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not explicitly state that source code for the described methodology is publicly available, nor does it provide a link. |
| Open Datasets | Yes | We evaluate i FAN on two domain adaptation scenarios: 1) train on SIM10K (Johnson-Roberson et al. 2017) and test on Cityscapes (Cordts et al. 2016) dataset (SIM10K Cityscapes); 2) train on Cityscapes (Cordts et al. 2016) and test on Foggy Cityscapes (Sakaridis, Dai, and Van Gool 2018) (Cityscapes Foggy). |
| Dataset Splits | Yes | The Cityscapes dataset has 3,475 images of 8 object categories taken from real urban scenes, where 2,975 images are used for training and the remaining 500 for evaluation. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU or CPU models, memory sizes) used for the experiments. |
| Software Dependencies | No | The paper mentions 'implement all with maskrcnn-benchmark (Massa and Girshick 2018)' but does not specify software versions for PyTorch, Python, or other key libraries. |
| Experiment Setup | Yes | The shorter side of training and test images are set to 600. The detector is first trained with a learning rate of lr = 0.001 for 50K iterations, and then lr = 0.0001 for another 20K iterations. The category-agnostic/aware instance-level alignment late launches at 30K-th iteration and categorycorrelation alignment at 50K-th. ... We set λadv = 0.1 in Eqn. 7 and λl = 1.0 in Eqn. 3. The embedding dimension of category-correlation alignment is set to 256, with a margin of m = 1.0. VGG-16 is used as the backbone if not specifically indicated. |