Age Label Distribution Estimation Algorithm Based on Kernel Extreme Learning Machine

Age label distribution learning can effectively describe the age label by adding information of adjacent ages. Based on this, this paper proposes an age label distribution estimation algorithm based on kernel extreme learning machine (ALDEA-KELM). Firstly, the features in the training and testing images are extracted, transform the age labels through the probability density function, convert age label to age label distribution. Secondly, establish an ALDEA-KELM regression model with features corresponding to the age label distribution. Finally, calculate the maximum probabilistic age and expected age of the predicted sample through the model. The maximum probability age and the expected age are combined for joint prediction. The comparison between other algorithms shows that method of this paper can effectively reduce the age prediction error. In terms of age prediction, the results come closer to the true values.