【摘要】 反映藏族文化艺术特色的古唐卡由于自然和人为的原因有些遭受了不同程度的损伤。针对唐卡画面颜色脱落的一类特殊损伤形式,通过对Hopfield神经网络稳定性的分析,在HSI色彩空间研究Hopfield网络模型,并利用其记忆特性对古唐卡图像进行颜色聚类分割。实验结果表明这种方法可以有效地提取图像中的损伤区域。更多还原

【Abstract】 The ancient Thangka image with reflecting Tibet art has been broken by the nature and human with different degree. The part-shedded Thangka image is researched as a type of special breaking and analyzed the stability of Hopfield Neural Network. The color clustering segmentation to broken Thangka image by researching Hopfield network model in the HSI color space and using it’s memory specialty can be performed. Finally the experiment result indicated that this method can extract effectively Thangka image breaking section.更多还原