Object detection and tracking for a vision guided automated suspended cell injection process

Object detection and tracking under the microscopic field of view is an important part in a vision based automated cell injection system. This paper introduces a novel method for recognizing and tracking the cell, the holding pipette and the injection pipette in an Intracytoplasmic Sperm Injection (ICSI) process. The algorithm for identification firstly automatically segmented the microscopic image by Otsu´s method, and then calculated the roundness and elongation features of every separated region after removing tiny blobs by judging their areas. The holding pipette was identified by a modified template matching technique, which used the contour information of the object instead of the whole body, resulting in a shorter processing time. The objects were tracked by applying identification method to every frame of the video, with a movement estimation constraint of the holding pipette when performing the matching. Experiment was carried out and the performance of the proposed method was evaluated.