Plasma cells, derived from B lymphocytes in the bone marrow, are important white blood cells that produce antibodies, helping combat bacteria and viruses and prevent inflammation.
The study presents a new model combining ResNet and UNet networks for effective segmentation and analysis of white blood cells (leukocytes) from blood samples, overcoming challenges in image processing.
Experimental results demonstrate that this model achieves approximately 96% segmentation accuracy, outperforming previous methods like DeepLabV3+ and ResNet-50 in diagnosing blood diseases.