Accurate and fast extraction of adhesive cells based on concave points detection and matching

Precise image segmentation of adhesive cells is a challenge during the cell detection process for biomedical image analyses. In this article, a new segmentation and extraction method for adhesive cells based on concave points detection and matching is proposed. First, image preprocessing is used to reduce interference information, which mainly includes sequential gray scale transformation, binarization, hole filling, adhesive cells recognition, and median filtering process. Second, an improved concave detection algorithm is presented which can reduce the interfering feature points and detect the concave points on the adhesive cells accurately. Third, precise extraction of adhesive cell contour is realized by concave matching based on ellipse fitting. Experiments of blood cell sample detection show that the average accuracy is 97.18% with a costing time of 3.36 s. The proposed method can effectively reduce the interference of multiple concave points and extract the cell contour from adhesive regions with higher accuracy and less time, which shows great application potential in the field of biological image detection and analyses.