Segmentation of White Blood Cells Based on CBAM-DC-UNet

Monitoring the morphology of blood leukocytes, plays an important role in medical research, especially treatment diseases such as immunodeficiency. Traditional manual detection methods are susceptible to numerous interference factors. Therefore, cells often segmented using deep-learning algorithms. This study proposes a U-Net model based on combination attention mechanism and dilated convolutions. First, traditional convolution double convolutional module network is replaced by convolution, multi-scale features obtained expanding receptive field. Second, after each layer upsampling layer, combined refine adaptive improve segmentation performance model. Finally, RAdam optimizer was used enhance robustness learning rate variations. Through ablation experiment three improvement directions, it concluded that all directions had positive effect result, most effective when improvements were combined. The experimental results show compared with original model, indicators intersection over union (IOU), recall accuracy increased 5.1%, 5.7% 1.2%, respectively, which more accurately may be for greater degree auxiliary leukocyte application immunodeficiency other diseases.