Pseudogenes are homologous relatives of known genes that have lost their ability to function as a transcriptional unit. Three classes of pseudogenes are known to exist: duplicated pseudogenes; processed or retrotransposed pseudogenes; and unitary or disabled pseudogenes. Since pseudogenes may display a number of the characteristics of functional genes, they pose a unique set of problems for ab initio gene prediction. The ability to detect and differentiate pseudogenes from functional genes can be a difficult task. We present a comprehensive review of current approaches for pseudogene detection, highlighting difficulties in pseudogene differentiation.