Arsenault H H
Appl Opt. 1989 May 1;28(9):1614-5. doi: 10.1364/AO.28.001614.
As Danielsson says in his comment on the paper by Arsenault, the phase information is not used and the same signature vectors can have different shapes. However, in practice the method works well.
正如丹尼尔松在对阿尔塞纳尔特论文的评论中所说,相位信息未被使用,相同的签名向量可能具有不同的形状。然而,在实际应用中该方法效果良好。