海量医学图像下的病理结果推定数学模型仿真

Mathematical Model Simulation of Pathological Results Presumption Based on Massive Medical Images

依据海量医学图像特征,对现有图像的病理结果进行推定时,由于海量医学图像数据量大,特征之间的关联性及其复杂,传统的算法进行病理结果推定,需要建立较为复杂的关联规则,造成推定计算效率较低,错误率较高。提出以大数据分析为基础的隐马尔科夫医学图像识别与病理结果推定方法。针对待测医学图像,采用双边滤波方法进行去噪处理,在去除图像数据噪声的同时,有效地保留了图像关键边缘的完整性,对边缘特征进行奇异值分解量化,减少不必要冗余特征干扰,根据隐马尔科夫原理计算图像的最大似然值,降低海量医学图像数据的病理结果推定的复杂程度。实验结果表明,利用改进算法进行基于海量医学图像数据的病理结果推定,能够提高计算效率与病理结果推定的准确性,提高推定效率。

In the paper, a hidden Markov medical image recognition and pathological results presumption method based on big data analysis was proposed. For the medical images to be detected, bilateral filtering method was used to make denoising. At the same time, in removing the noises of image data, it effectively retained the integrity of the key image edge, made the singular value decomposition quantification of the edge feature, and reduced unnecessary re- dundancy feature interference. The maximum likelihood value of images can be computed according to the principle of hidden Markov, to reduce the complexity of the pathological results presumption of massive medical image data. Ex- perimental results show that using the improved algorithm for the pathological results presumption of massive medical image data can improve the accuracy of the calculation efficiency, the pathological results presumption, and the effi- ciency of presumption.