EfficientRF: Facial Age Estimation Based on EfficientNet and Random Forest

Age estimation by analyzing facial images has always been a challenging problem in computer vision. With the maturity of face recognition technology and the popularity of mobile devices, it is of great significance to deploy face age estimation algorithms on mobile devices. Aiming at the problems of high model complexity and low running speed of the current face age estimation model, this paper proposes a new end-to-end model named EfficientRF for the age estimation task. EfficientRF integrates random forest based on EfficientNet network architecture, wherein the split nodes of the random forest are connected to the last fully connected layer of the convolutional neural network. The experimental results show that, on the premise of ensuring high accuracy, the new method uses 1.6 times less parameters and 60 times faster running speed in face age estimation compared to three public age estimation benchmarks, and is superior to the most advanced standard on FG-NET dataset.