论文部分内容阅读
在试验设计方法学的基础上，提出了一种定量评价自动目标识别算法性能的方法；给出了几种刻划场景变化的图像质量指标的定义和度量方法；提出了用响应曲面模型方法建立算法性能模型和描述图像度量与算法性能之间统计关系的方法．实验结果表明，仅需要少量的测试图像就可得到可靠的算法性能模型和可信的评价结果．
Based on the experimental design methodology, a method of quantitatively evaluating the performance of automatic target recognition algorithm is proposed. Several definition and measurement methods of image quality indicators are given. A method of response surface modeling is proposed Algorithmic performance models and methods to describe the statistical relationship between image metrics and algorithm performance. The experimental results show that only a small number of test images can be obtained reliable algorithm performance model and credible evaluation results.