Integrated Age Estimation Mechanism based on Decision-Level Fusion of Error and Deviation Orientation Model

Abstract

Age estimation based on machine learning has received lots of attention. Traditional age estimation mechanism focuses age error ignoring the deviation between the estimated age and real age due to disease. Pathological age estimation mechanism used age deviation as the training label to solve the above problem. However, it results in a larger error between the estimated age and real age in the normal control (NC) group. An integrated age estimation mechanism based on Decision-Level fusion of error and deviation orientation model is proposed to solve the problem. Firstly, the traditional age and pathological age estimation mechanisms are weighted together. Then, their optimal weights are obtained by minimizing mean absolute error (MAE). Finally, with the optimal weight, the integrated age estimation mechanism (IAE) is built. Several representative age-related datasets are used for verification. The results show that the proposed age estimation mechanism achieves a good tradeoff effect of age estimation. It not only improves the classification ability of the estimated age, but also reduces the age estimation error of the NC group.