yuantn / MI-AOD

Code for Multiple Instance Active Learning for Object Detection, CVPR 2021
https://openaccess.thecvf.com/content/CVPR2021/papers/Yuan_Multiple_Instance_Active_Learning_for_Object_Detection_CVPR_2021_paper.pdf
Apache License 2.0
333 stars 43 forks source link

能否只使用代码中的主动学习部分来筛选图片呢 #95

Closed dust-removal closed 1 year ago

dust-removal commented 1 year ago

你好,我刚看完你的论文,我最近有一个应用场景需要使用主动学习,我想知道能够用你的方法仅仅帮我筛选图片,应该怎么做。

yuantn commented 1 year ago

你可以修改 mmdet/utils/active_datasets.py 文件中的 update_X_L 函数(它在 tools/train.py 中被调用),令其返回 X_S,此即为每一轮主动学习后筛选出不确定性较高的新增样本。


You can modify the update_X_L function in mmdet/utils/active_datasets.py called by tools/train.py, make it return X_S, which are the new selected samples with higher uncertainty after each cycle of active learning.

dust-removal commented 1 year ago

你可以修改 mmdet/utils/active_datasets.py 文件中的 update_X_L 函数(它在 tools/train.py 中被调用),令其返回 X_S,此即为每一轮主动学习后筛选出不确定性较高的新增样本。

You can modify the update_X_L function in mmdet/utils/active_datasets.py called by tools/train.py, make it return X_S, which are the new selected samples with higher uncertainty after each cycle of active learning.

感谢,我现在发现mmdet2.3.0已经有了支持windows的版本,那是不是在win上运行就有可能了

yuantn commented 1 year ago

是的,但我没有尝试过,你可以试一下。


Yes, but I have not tried it. You may have a try.