The traditional gray-level co-occurrence matrix (GLCM) was computationally intensive and discriminatively insufficient.

分析了传统的灰度共生矩阵在计算纹理特征时计算量大，且分辨能力差的缺点。