An automatic method for red blood cells detection in urine sediment micrograph

Urine sediment micrograph consists of various tangible components, such as red blood cells (RBCS), white blood cells (WBCs), tube and crystal, etc. Quantitative analysis of urine sediment micrograph is of great significance for infectious diseases and circulatory diseases diagnosis. The traditional method about urine sediment analysis depends on the observation of medical staff, in that case the workload is huge. With the development of image processing and pattern recognition techniques, the automation of urine sediment analysis can be realized. However, due to the complexity of the urine sediment micrograph, the accuracy and efficiency for automatic analysis are still in a low level somewhat. In this paper, an automatic detection method is proposed for the RBCs in the urine sediment micrograph. We borrow the concept of channel features which contain diverse type color channel features, and gradient magnitude channel features, etc. We adopt aggregate channel features which are variant and discriminative, combing improved soft-cascade adaboost classifier for RBCs detection in urine sediment micrograph. On collected challenging dataset, it shows competitive performance compared with Support Vector Machine (SVM) using Histogram of Oriented Gradient (HOG).