CELL RECOGNITION BY IMAGE-PROCESSING - RECOGNITION OF DEAD OR LIVING PLANT-CELLS BY NEURAL-NETWORK

This paper deals with determination of whether protoplasts are alive o r not. The protoplasts are plant cells which are removed cell walls, a nd have been used for cell fusion to generate new species of plants. C ell fusion uses living protoplasts, and therefore, a method of life or death determination for protoplasts is necessary. FDA (fluorescein di acetate) has been used for this purpose. However, FDA has not been app lied for cell fusion because it is toxic to protoplasts. In this paper , we propose a new harmless method for the above determination using t he image processing technique and an artificial neural network. The pr oposed method measures shape and colors of protoplasts using image pro cessing for sensory information. The neural network determines whether protoplasts are living or not based on the sensory information. Exper iments were performed in order to show the effectiveness of the propos ed method. As the result, the proposed method is found to be able to r ecognize states of protoplasts with an accuracy of 85 percent. This me thod is promising for determining life or death of protoplasts in cell fusion, instead of FDA.