anshumaan031 / detection-of-blood-cells-using-yolov5

Detection-of-Blood-Cells-using-Yolov5

Blood is the only fluid connective tissue, which functions by circulating and transporting oxygen, nutrients and other essential minerals to various cells and tissues of our body.

In humans, blood is mainly composed of plasma, blood cells and platelets. Overall, blood makes up 7 to 8 percent of total body weight and an average, healthy person possesses around 5 to 6 litres of blood.
YOLOv5 algorithm was used for this project to detect blood cells like Red Blood Cells, White Blood Cells and Plateletes. Object detection is the task of classifying as well as localizing the objects in an image. In yolo, the entire image is divided into grids with each of them having their own vectors representing the coordinates of bounding box and class type. LabelImg software was used to manually annotate the image as well as define the bounding box. The images and the XML files conatining the bouding box info are created as a result to train our model with custom weights.The darknet repository was cloned into the notebook to apply the object detection algorithm on our custom dataset.