| 查看: 231 | 回复: 1 | |||
| 当前主题已经存档。 | |||
[交流]
求助翻译 修改
|
|||
|
随着计算机处理速度的不断提高，计算机模拟人类视觉方面取得长足进展，在图像识别与分析系统中采用贝叶斯决策和半监督[5]的方法已经取得较好效果，如用于字体的判断等。本文以织物图像分割为例，提出了用计算机图像处理技术与以最小风险贝叶斯决策理论为判断准则的半监督聚类理论相融合的图像分割算法。 With the improvement of computer’s processing speed, we have made great advances in computer simulation of human vision, Bayesian decision-making and semi-supervised method has been achieved good results in image recognition and analysis system, such as determine of fonts and so on. This thesis takes image segmentation for a case, presents a image segmentation algorithm that combine computer image processing technology with the theory of semi-supervised clustering, which gives the judgement criterion based on minimize risk of Bayesian decision theory. 通常，聚类算法通过一个优化准则函数来确定对数据的 个划分。但是寻找分割问题的最优解是一个 难的问题。在机器学习领域，分类属于监督学习。绝大多数的有监督的机器学习方法依赖于标注的训练样本集，忽略了未标注样本的作用，利用大规模的标注过的训练数据固然可以提高学习算法结果的准确度，但是标记必须由人工完成，这是一项费时费力的工作。基于此，提出了半监督的聚类算法，聚类是图像分割的重要方法，半监督学习算法就是利用这些未标注样本，在传统的机器学习方法中结合未标注样本进行学习的算法，半监督的思想可以很好地将先验的分割信息融合到图像的分割过程中。 Generally, the clustering algorithm adopts a criterion function of optimization to divide data into K groups. However, the problem of finding the optimal solution of segmentation is NP-hard. In the field of machine learning, classification is supervised learning. Most supervised machine learning methods rely on tagging the training sample set, ignores the role of non-marked samples, uses large-scale training data of tagging can improve the accuracy of the result of learning algorithm, but the tags must be finished by hand, which is a hard, lengthy work. Based on this, this paper puts forward a semi-supervised clustering algorithm, clustering is an important method for image segmentation. Semi-supervised learning algorithm is an algorithm that combines non-marked samples with traditional machine learning methods. By the mean of semi-supervised, a priori segmentation information will be fused to the process of image segmentation as well. 请高手帮忙 ，，把这2段的语法和句子结构 修改下。。 |
回复此楼» 猜你喜欢
085600材料与化工301分求调剂院校
已经有5人回复
-
081700，311，求调剂
已经有15人回复
-
一志愿北京化工085600 310分求调剂
已经有18人回复
-
材料与化工371求调剂
已经有14人回复
-
336材料与化工085600求调剂
已经有7人回复
-
材料334求调剂
已经有18人回复
-
331求调剂
已经有8人回复
-
332求调剂
已经有17人回复
-
一志愿南京航空航天大学 材料与化工329分求调剂
已经有4人回复
-
材料专硕322
已经有7人回复
jijun2003860
铜虫 (小有名气)
- 应助: 0 (幼儿园)
- 金币: 138.5
- 红花: 1
- 帖子: 84
- 在线: 19.2小时
- 虫号: 540962
- 注册: 2008-04-07
- 专业: 资源化工
2楼2009-04-15 10:13:49
