White Blood Cell Detection Based on FBDM-YOLOv8s

White blood cells, also known as immune cells, help the body resist infectious diseases and foreign pathogens as part of the immune system. This article proposes a white blood cell detection method based on an improved YOLOv8s to enhance the accurate detection of different categories of normal white blood cells. In the YOLOv8s network, this study replaces the main framework of YOLOv8 with the Fasternet module to reduce computational redundancy and memory access. Subsequently, a BiLevel Routing Attention mechanism is proposed to achieve precise localization of white blood cells. Additionally, to attain diversified feature extraction, this article introduces the DBBNCSPELAN4 module, which can perform diversified feature extraction in a multi-branch manner while lightweighting the model, reducing the number of parameters, and speeding up computation. Finally, the MultiSEAM module is integrated with Detect to address detection errors caused by occlusion and connected white blood cells. Experiments were conducted on a dataset of five types of normal cells provided by DML-LZWH (Liuzhou Workers’ Hospital Medical Laboratory). The FBDM-YOLOv8s significantly outperforms other advanced object detection algorithms in terms of performance, with a 2.1% increase in mAP compared to the baseline YOLOv8s.We will release the source code on <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/SSRR-LLL/FBDM-YOLOv8.git</uri>.