Popularity of plastic surgery has increased many folds over the past
few years and the statistical data shows that it keeps growing. Due to
advances in technology, affordability, and the speed with which these
procedures can be performed, several people undergo plastic surgery
for medical reasons and some choose cosmetic surgery to look
younger or for better appearance. The procedures can significantly
change the facial regions both locally and globally, altering the
appearance, facial features and texture, thereby posing a serious
challenge to face recognition systems. Existing face recognition
algorithms generally rely on local and global facial features and
any variation can affect the recognition performance. This paper
introduces plastic surgery as a new dimension for face recognition
algorithms. We present an experimental study to quantitatively evaluate
the performance of face recognition algorithms on a plastic
surgery database that contains face images with both local and global
surgeries. The study shows that appearance, feature, and texture based
algorithms are unable to effectively mitigate the variations caused
by the plastic surgery procedures. Based on the results, we believe
that more research is required in order to design an optimal face
recognition algorithm that can also account for the challenges due
to plastic surgery. It is our assertion that the results of this work
would inspire further research in this important area. One possible
future research direction would be to use thermal-infrared imagery
and compute the thermal differences between pre and post surgery
images. However, such an approach first requires creating a large
face database that contains pre and post operative thermal infrared
images.