This work proposes an intelligent visual technique for detecting phishing websites. The phishing websites are classified into three categories: very similar, local similar, and non-imitating. For cases of ‘very similar’, this study uses the wavelet Hashing (wHash) mechanism with a color histogram to evaluate the similarity. In cases of ‘local similarity’, this study uses the Scale-Invariant Feature Transform (SIFT) technique to evaluate the similarity. This work concerns ‘very similar’ and ‘local similar’ cases to detect phishing websites. The results of the experiments reveal that the wHash mechanism with a color histogram is more accurate than the currently used perceptual Hashing (pHash) mechanism. The accuracies of SIFT technique are 97.93%, 98.61%, and 99.95% related to Microsoft, Dropbox, and Bank of America data, respectively.

中文翻译：

---

基于智能视觉相似性的钓鱼网站检测

这项工作提出了一种用于检测网络钓鱼网站的智能视觉技术。钓鱼网站分为三类：非常相似、本地相似和非模仿。对于“非常相似”的情况，本研究使用带有颜色直方图的小波哈希（wHash）机制来评估相似度。在“局部相似性”的情况下，本研究使用尺度不变特征变换 (SIFT) 技术来评估相似性。这项工作涉及检测网络钓鱼网站的“非常相似”和“本地相似”案例。实验结果表明，带有颜色直方图的 wHash 机制比目前使用的感知哈希（pHash）机制更准确。SIFT 技术与 Microsoft、Dropbox 和美国银行数据相关的准确率分别为 97.93%、98.61% 和 99.95%。