摘要
分割是自动目标识别系统(ATRS)的关键部分,其结果直接影响着其后的信息处理过程.评估分割算法的性能是ATRS离线评估的重要组成之一,本文提出了一个确定分割算法性能指标的规则,并据此建立了一种新的依赖于目标在图像上的面积因素和形状因素的指标计算方法,应用实例显示出它对于分割算法的性能评估比其它的一些方法更为合理.
Segmentation is the key component in automatic target recognition systems. Its results have critical implications for the information processing stragies that follow. The evaluation of a segmentation algorithm takes an important role in the estimation of performance for an off-line ATRS. In this paper, some criteria to determin metrics for the evaluation of segmentation algorithms are presented. According to these criteria, a new method to calculate the metrics based on the area factors and form factors of targets on an image is established. Its application shows that the new method is more reasonable than other methods.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
1996年第2期144-148,共5页
Pattern Recognition and Artificial Intelligence
基金
国家自然科学基金
关键词
自动目标识别
性能评估
图像识别
图像分割
Automatic Target Recognition, Segmentation, Performance Evaluation.
