Weight mechanism: adding a constant in concatenation of series connect
This work addresses feature fusion issues in image segmentation for building detection, but it is incremental as it builds on existing U-Net architectures with minor modifications.
The paper tackles the problem of semantic deviation and noise in feature fusion by proposing a weight mechanism to adjust concatenation in series connections, resulting in a 0.80% mIoU improvement on the Massachusetts building dataset and a 0.12% mIoU gain in a fused U-Net architecture.
It is a consensus that feature maps in the shallow layer are more related to image attributes such as texture and shape, whereas abstract semantic representation exists in the deep layer. Meanwhile, some image information will be lost in the process of the convolution operation. Naturally, the direct method is combining them together to gain lost detailed information through concatenation or adding. In fact, the image representation flowed in feature fusion can not match with the semantic representation completely, and the semantic deviation in different layers also destroy the information purification, that leads to useless information being mixed into the fusion layers. Therefore, it is crucial to narrow the gap among the fused layers and reduce the impact of noises during fusion. In this paper, we propose a method named weight mechanism to reduce the gap between feature maps in concatenation of series connection, and we get a better result of 0.80% mIoU improvement on Massachusetts building dataset by changing the weight of the concatenation of series connection in residual U-Net. Specifically, we design a new architecture named fused U-Net to test weight mechanism, and it also gains 0.12% mIoU improvement.