基于坎贝拉度量改进的智能拼图算法  

Jigsaw Puzzles Based on Canberra Distance

作　　者：吴娟 WU Juan(College of Engineering professional,Jinan University,Jinan 250200,Shandong)

机构地区：[1]济南工程职业技术学院信息工程学院,山东济南250200

出　　处：《电脑与电信》2022年第8期64-68,共5页Computer & Telecommunication

摘　　要：智能拼图是根据相邻块的相似度进行组装的,因此如何计算拼图块之间的相似度度量至关重要。提出了一种新颖的匹配拼图块的相似度度量方法,考虑相邻拼图块之间边缘的相似度,将坎贝拉距离与MGC算法(马氏相似度度量)结合,确定阈值后利用贪婪算法将拼图块进行组装。将部分网络图片以及现实拍摄的部分图片进行测试,实验结果表明本文算法具有更高的准确率。As puzzles are assembled according to the similarity of adjacent blocks,therefore how to calculate pairwise compatibility is crucial.This paper proposes a new method that considers edge similarity between puzzle pieces and combines the Canberra distance with the MGC algorithm.After making the threshold,the greedy algorithm is used to assemble the puzzle pieces through the combination of MGC.Some network pictures and realistic pictures are tested.The experimental results show that the algorithm has higher accuracy.

关 键 词：拼图 坎贝拉距离 MGC(马氏相似度度量) 贪婪算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]