Lung image segmentation based on Involution UNet model
Hongxin Xiao, Lingxi Peng, Shaohu Peng, Yifan Zhang · 2022 5th International Conference on Advanced Electronic Materials, Computers and Software Engineering (AEMCSE) · 2022
The existing lung segmentation algorithms have problems such as insufficient feature extraction ability and low segmentation efficiency. To address this problem, we propose a new UNet-like model called In-UNet. In order to expand the perceptual field and reduce feature loss, we integrated the involution and Inception modules into U-Net. In addition, the segmentation efficiency of the network was improved by using the Mish activation function. The experimental results on the lung CT dataset show that the method can segment the lung parenchyma region effectively. It also has better segmentation effect than other segmentation networks.