尺度不变特征变换算子综述（"计算机视觉前沿论坛"专栏征文）

随着计算机软件与硬件技术的发展，计算机视觉算法逐渐成为图像处理领域的研究热点。其中SIFT（scale invariant feature transform）算法是目前机器视觉领域应用最成功的算法之一。由于在尺度不变、旋转不变、光照不变等方面的独特优势，SIFT被广大视觉领域的研究者借鉴与学习。但是SIFT算法本身也存在一些问题，如仿射性能不太理想、计算复杂度过高等，因此针对它的多种改进算法不断出现。本文对SIFT的发展历史、SIFT算法的演变以及它不同领域的典型应用给出了一个比较全面的综述，并比较了各类算法的优缺点。最后给出了该算法未来可能的发展方向，为视觉研究者提供参考。

Summarization of Scale Invariant Feature Transform

With the development of software and hardware technique, the computer vision has become the hot research fields in image processing. SIFT (scale invariant feature transform) is one of the most successful vision algorithm nowadays and it is studied by the vision researcher widely because of its unique features in scale invariant, rotation invariant and illumination invariant. However, it also has some problems such as only part affine and the high computation complexity. Many extended or modified algorithms of the SIFT are developed unceasingly. The paper summarizes the develop history, the evolved processing and the application of the SIFT and compares those algorithm effect. At last the paper discusses the feature direction and provides reference for vision researchers.