The Recognition Of The Cervical Smear Based On Artificial Neural Network

At present, digital picture processing, the pattern recognition and the artificial intelligence technology already widely was applied to the biology medicine domain. Some encouraging results, the medicine picture processing system report which diagnosed about the cancer early time pathology are not many. The cancer cell's examination and the diagnosis was still the important and one difficult work of the medical worker. This article carried on the exploration in the antithetical couplet cervical cell examination and the diagnosis aspect, and obtained the encouraging result.First, to gains cervical the picture carries on the gradation transformation. Transforms by the original 24-bit colored pictures as the gradation picture. In carries on the division to the gradation picture, mainly adopts based on the threshold value division method. Separately to the cell, the cell nucleus carried on the division. The after division transforms into two values pictures, uses eight to includes the perimeter to the chain code algorithm, the area resembles the roundness, rectangular, corn compared to and so on ,15 main morphology parameters is carries on the survey. After obtained the massive data sample carries on the nerve network the training.In this article the nerve network model mainly is the BP reverse propagated error algorithm. First front the application obtains the massive data sample carries on the training to the nerve network power value. After the error is smaller than the rating, carries on the test in the application data sample to the nerve network. May know through the massive experimental contrast result, may the antithetical couplet cervical cell carry on a more accurate classification using the BP reverse propagated error algorithm. Obviously BP reverse propagated error algorithm quite good classification ability.Needs unifies in this system research performance history the computer technology and the pathology expert's practical experience, in uses the picture processing technology processing picture in the foundation, carries on the classified recognition using the nerve network to the cell. This has the certain practical significance in the medicine research as well as the clinical diagnosis aspect and the quite broad application background.