The hybrid operation combination of the micro and nano manipulation plays a crucial role in studying living cells’ biological process and the interactions of cells and drugs. A method is presented to monitor and analyze tumor cells’ biological process by image processing techniques based on an optical microscope in this paper.This paper described research status in the domain of image recognition. In the bone marrow image segmentation, an improved image segmentation algorithm is presented. An improved line connection algorithm and improved hole filling algorithm have been proposed to complete finally segmentation. Next, Features are extracted based on image segmentation, concave spot detection is useful to recognize abnormal cells. At last, living tumor cells image recognition is proposed based on BP neural network. Meanwhile, cells are counted based on connected component labeling algorithm.Experimental results demonstrate that the elapsed time of the whole recognition process is reasonable, and the recognition rate is high. |