基于天牛须优化的BAS-CRI图像修复算法

BAS-CRI Image Inpainting Algorithm Based on the Optimization of Beetle Antennase Search Algorithm

针对Criminisi算法中由于优先权计算不准确导致的修复效果不理想、修复效率低等问题,提出一种基于天牛须算法优化的BAS-CRI图像修复算法.首先在Criminisi算法优先权计算过程中,加入等照度曲线提高优先权的可靠性,并针对置信项加入调整因子,平滑置信项.其次,在最佳匹配块搜索阶段,引入改进后的天牛须搜索算法,提高搜索效率和鲁棒性.最后,利用主观视觉评价和峰值信噪比(PSNR值)量化评价,验证了算法的有效性.结果表明,所提算法较其他算法修复效果更好.

In order to solve the problems of poor repair effect and low repair efficiency in Criminisi algorithm,caused by inaccurate priority calculation. A new intellgent algorithm BAS-CRI image restoration algorithm based on the optimization of the beetle antennase search algorithm is proposed. Firstly,in the priority calculation process of Criminisi algorithm,the equal illuminance curve is added to improve the reliability of priority. Then,adding the adjustment factor to the confidence term to smooth the confidence item. Moreover,when searching the optimal matching blocks,the beetle antennase search algorithm is introduced to improve the search efficiency and robustness. Finally,the effectiveness of the proposed algorithm is verified by subjective and quantitative evaluation through visual evaluation and peak signal-tonoise ratio( PSNR),and the proposed algorithm is better than other algorithms.