Similar fragments in source codes are known as clones or duplicated codes. One major issue with duplicated fragments is that occurring a bug in one code makes all similar codes buggy. This requires the software developer to check all similar codes to fix the bug. As a result, the cost of software development when using available source codes may increase. Therefore, detecting similar codes seems to be inevitable not only for fixing the bugs, but also for copyright purposes. In this paper we briefly review clone detection techniques especially graph-based approacheswhich detect exact and approximate clones in source codes.