Quality evaluation of facial image toward age estimation

The quality of human facial image has a strong influence on the performance of age estimation by facial images. Until now, few efforts directly focus on quality evaluation of facial image for age estimation. This paper tries to solve this gap. Our work includes: (1) Propose two types of evaluation indicators, including general evaluation indicators and special evaluation indicators. (2) Give the computational evaluation methods to get the indicators from facial images. To evaluate our work, some images from Album 1 of Morph database are selected to form the training and testing sets. Experimental results show that the proposed method can effectively evaluate the quality of the facial images.