CAM-SegNet: A Context-Aware Dense Material Segmentation Network for Sparsely Labelled Datasets
CAM-SegNet: A Context-Aware Dense Material Segmentation Network for Sparsely Labelled Datasets
复制标题
DOI：
10.5220/0010853200003124
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Yuwen Heng;Yihong Wu;S. Dasmahapatra;Hansung Kim
中科院分区：
文献类型：
--
作者：
Yuwen Heng;Yihong Wu;S. Dasmahapatra;Hansung Kim
上下文信息减少了密集材料分割任务中的不确定性，以提高分割质量。典型的上下文信息包括对象、地点标签或通过神经网络提取的特征图。现有方法通常采用预先训练的网络来生成上下文特征图，而无需微调，因为专用材料数据集不包含上下文标签。因此，这些上下文特征可能不会改善材料分割性能。考虑到这个问题，本文提出了一种混合网络架构，CAM-SegNet，在训练过程中联合学习上下文和材料特征，而无需额外的上下文标签。我们的CAM-SegNet的实用性通过引导网络在自我训练方法的帮助下学习边界相关的上下文特征来证明。实验表明，CAM-SegNet可以识别具有相似外观的材料，准确率提高3-20%，平均IoU提高6-28%。
: Contextual information reduces the uncertainty in the dense material segmentation task to improve segmentation quality. Typical contextual information includes object, place labels or extracted feature maps by a neural network. Existing methods typically adopt a pre-trained network to generate contextual feature maps without fine-tuning since dedicated material datasets do not contain contextual labels. As a consequence, these contextual features may not improve the material segmentation performance. In consideration of this problem, this paper proposes a hybrid network architecture, the CAM-SegNet, to learn from contextual and material features during training jointly without extra contextual labels. The utility of our CAM-SegNet is demonstrated by guiding the network to learn boundary-related contextual features with the help of a self-training approach. Experiments show that CAM-SegNet can recognise materials that have similar appearances, achieving an improvement of 3-20% on accuracy and 6-28% on Mean IoU.