自动目标识别算法性能评估中的图像度量

图像度量是自动目标识别(ATR)性能评估中的重要组成部分。图像度量是否与ATR算法性能紧密相关将直接影响系统后续的评价工作。先介绍了传统的图像度量，并分析了作为传统度量代表的目标局部背景对比度度量(TBC)在复杂场景条件下与算法性能不满足单调关系的不足，针对其局限性，提出了基于灰度共生矩阵的图像杂波度量(TIC)，并针对TBC和TIC设计了两组实验。结果表明，无论在指定场景还是复杂场景条件下，TIC 与算法性能都具有良好的单调关系,有效地克服了TBC的局限性，从而能更好地评价ATR算法性能．

Image measures in the evaluation of ATR algorithm performance

Image measure is a very important part of automatic target recognition algorithm performance evaluation. Whether the image measure is tightly related with algorithm performance affects directly the evaluation works. The current image measures are introduced, and the deficiency in the target to background contrast(TBC) measure which is the representative of current general measures is analyzed. A new texture-based image clutter measure(TIC) is proposed to solve the deficiency in TBC. Experimental result of testing two measures TBC and TIC shows that TIC has very good monotonic relation with the segmentation algorithm performance in both of the given conditions,which proves TIC as a robust indicator of segmentation algorithm performance and gives better performance than TBC.