Particle Detection in Intracellular Images and Radius Estimation by Circle Fitting

Detecting particles in intracellular images and estimating their radii is important for investigating the causes of clinical conditions. Presently, particles are manually counted by observers, which is time consuming and subjective. Thus, we propose an automatic particle counting and radius estimation method based on pattern recognition techniques. Particles are analyzed by a computer in two processes. During the first process, particles are detected using face recognition via a support vector machine. In the second process, the edge points of a detected particle are distinguished by the intensity difference between the particle and the background. The edge points are extracted and fitted by a circle by a voting technique. The proposed method accurately located 96.04% of the particles detected by manual counting. ImageJ, the particle detection software generally used in cell biology, detected 72.32% of the particles. Therefore, our method yields higher accuracy than ImageJ. Moreover, the effectiveness of radius estimation was also experimentally verified. (c) 2014 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.