arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.00875cs.CV
收藏

类增量式语义分割的半真半假提取

Half-Real Half-Fake Distillation for Class-Incremental Semantic Segmentation

Huazhong University of Science and Technology(华中科技大学);Alibaba Group(阿里巴巴集团)

Zilong Huang, Wentian Hao, Xinggang Wang, Mingyuan Tao, Jianqiang Huang, Wenyu Liu, Xian-Sheng Hua

摘要:

尽管它们在语义分割方面取得了成功,但卷积神经网络对于增量学习\ie来说装备不足,当新的类可用时,卷积神经网络调整原始分割模型,但不保留初始训练数据。事实上,他们很容易出现灾难性遗忘问题。我们试图通过“反转”训练的分割网络来解决这个问题,以从随机噪声开始合成输入图像。为了避免手动设置详细的逐像素分割图作为监督,我们提出了SegInversion使用图像级标签来合成图像。为了增加合成图像的多样性,Scale-Aware Aggregation模块集成到SegInversion中,用于控制合成对象的比例(像素数)。与新类别的真实图像一起,合成的图像将被馈送到基于蒸馏的框架中,以训练新的分割模型,该模型保留关于先前学习的类别的信息,同时更新当前模型以学习新的模型。该方法显著优于其他增量学习方法,并在PASCAL VOC 2012和ADE 20 K数据集上获得了最先进的性能。代码和模型将公开提供。

英文摘要:

Despite their success for semantic segmentation, convolutional neural networks are ill-equipped for incremental learning, \ie, adapting the original segmentation model as new classes are available but the initial training data is not retained. Actually, they are vulnerable to catastrophic forgetting problem. We try to address this issue by "inverting" the trained segmentation network to synthesize input images starting from random noise. To avoid setting detailed pixel-wise segmentation maps as the supervision manually, we propose the SegInversion to synthesize images using the image-level labels. To increase the diversity of synthetic images, the Scale-Aware Aggregation module is integrated into SegInversion for controlling the scale (the number of pixels) of synthetic objects. Along with real images of new classes, the synthesized images will be fed into the distillation-based framework to train the new segmentation model which retains the information about previously learned classes, whilst updating the current model to learn the new ones. The proposed method significantly outperforms other incremental learning methods and obtains state-of-the-art performance on the PASCAL VOC 2012 and ADE20K datasets. The code and models will be made publicly available.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑