ID 原文 译文
47906 在噪声检测部分,由像素二阶差对各个子窗口像素进行像素边界值检测,进而对所有子窗口边界值进行全局统计确定图像噪声像素边界, In noise testing section, for each child window pixel by pixel second-order difference is used to detect the boundary value of pixels, and then to all child window boundary values for global statistics to determine the image boundary noise pixel,
47907 并利用同类像素个数对噪声边界内的像素进行纠正以降低错检率。 and use the same number of noise within the boundary pixels of the pixel correction in order to reduce the rate of fault detection.
47908 针对噪声像素,利用图像规则函数和噪声像素限制条件来构造规则函数滤波器。 In view of the noise pixels using the rules of image function and noise pixels rules limiting condition to construct a function filter.
47909 将所提出的方法应用于噪声像素检测与滤波,并与其他算法对比, The proposed method was applied to detection and filtering the noise pixel, and compared with other algorithms,
47910 实验结果表明,该方法能够同时保证低漏检率和错检率; the experiment results show that this method can guarantee lower miss rate and the fault detection rate at the same time;
47911 在滤波方面,该方法所得到的修复图像具有更高的峰值信噪比和视觉效果。 In terms of filter, the method of restoration image has higher peak signal to noise ratio and visual effect.
47912 针对小样本条件下的离散贝叶斯网络参数学习问题,提出一种基于单调性约束的学习算法。 For discrete bayesian network parameters under the condition of small sample learning problems, put forward a learning algorithm based on monotone constraint.
47913 首先,给出了单调性约束的数学模型,以表达定性的先验信息; First of all, gives the mathematical model of the monotonicity constraints, to express the qualitative prior information;
47914 然后,将单调性约束以狄利克雷先验的形式集成到贝叶斯估计中,并利用贝叶斯估计进行参数学习; Then monotonicity constraints in the form of dirichlet prior is integrated into bayesian estimation, and the use of bayesian estimation for parameter learning;
47915 最后,通过仿真实验与最大似然估计和保序回归方法进行比较。 Finally, the simulation experiment and maximum likelihood estimation and isotonic regression method is used in the comparison.