Talk:Facial Age Estimation

The following two important works are not cited.

Yun Fu, Guodong Guo and Thomas S. Huang, Age Synthesis and Estimation via Faces: A Survey, IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2010. (in press)

Yun Fu and Thomas S. Huang, Human Age Estimation with Regression on Discriminative Aging Manifold, IEEE Transactions on Multimedia (T-MM), Vol. 10, No. 4, pp: 578-584, 2008.

Some efforts should be made to distinguish the age estimation techniques from the age simulation techniques (a.k.a. age progression techniques). They are inverse processes to each other. While most early work about aging progress is on age simulation, age estimation is more recently studied.

Some basic ground-truth in the area of Physiology about aging progression might be helpful. For example, facial changes during early ages are mainly related to the craniofacial growth, while those during adult ages are mainly related to skin aging (texture changes).

Criteria to measure the performance of an age estimator should be introduced. This includes the Mean Absolute Error (MAE) and the cumulative score (CS), as mentioned in “X. Geng, Z. Zhou, Y. Zhng, G. Li, and H. Dai. Learning From facial aging patterns for automatic age estimation. Proceedings of ACMM, pp.307-316, 2006”.

Typos: 1. “Problem Definition”: 1st paragraph, “humans are no so accurate” -> “humans are not so accurate”. 2. “Problem Definition”: 2nd paragraph, “an age estimation system is indented to be used” -> “an age estimation system is intended to be used”