|Institution:||KTH Royal Institute of Technology|
|Keywords:||Natural Sciences; Computer and Information Science; Computer Science; Naturvetenskap; Data- och informationsvetenskap; Datavetenskap (datalogi)|
|Full text PDF:||http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-166771|
Classifying images and deciding algorithmically whether or not they are the same image is a complex problem in com-puter science. This report looks at using perceptual hashing in order to fingerprint images so that they may be matched to one and other. The results shows that threshold match-ing images can be done with good accuracy, while a zero distance match is not reliable in most scenarios. The con-clusion is that a hybrid approach that uses a more inclu-sive perception algorithm to filter out candidates for a more thorough algorithm might be a good choice.