基于感知哈希的作业相似度检测

Perceptual Hashing Based Plagiarism Detection of Homework

现有抄袭检测多以文本的字符串匹配实现,无法检测图片作业,有一定局限性。提出一种将文本作业转化为图片,应用感知哈希算法获取图片特征值并计算图片相似度的方法,实现作业的抄袭检测。实验表明,该方法在一定的检测速度下,能保证低误判率和查准率。

Plagiarism detection systems are mainly detecting by matching the string which causes restrictions, forinstance, image homework doesn＇t apply to these systems. This paper achieves the detection by using the per-ceptual hashing algorithm to get features of the images transformed from homework and calculating the similarityof these features. The experiments show that the method keeps the detection speed while ensures precision rateand false positive rate.