Robust Face Alignment with Random Forest: Analysis of Initialization, Landmarks Regression, and Shape Regularization Methods

Abstract

SUMMARYRandom forest regressor has recently been proposed as a local landmark estimator in the face alignment problem. It has been shown that random forest regressor can achieve accurate, fast, and robust performance when coupled with a global face-shape regularizer. In this paper, we extend this approach and propose a new Local Forest Classification and Regression (LFCR) framework in order to handle face images with large yaw angles. Specifically, the LFCR has an additional classification step prior to the regression step. Our experiment results show that this additional classification step is useful in rejecting outliers prior to the regression step, thus improving the face alignment results. We also analyze each system component through detailed experiments. In addition to the selection of feature descriptors and several important tuning parameters of the random forest regressor, we examine different initialization and shape regularization processes. We compare our best outcomes to the state-of-the-art system and show that our method outperforms other parametric shape-fitting approaches. key words: face alignment, shape model fitting, random forest, point distribution model, sparse representation model

Real-time face alignment: evaluation methods, training strategies and implementation optimization

“…There are two type of approaches: local and global approaches depending on using global face appearance or local patches. Typical methods inside of this category are the Regression Forests methods [34], simple and low computational complexity algorithms such as the Conditional Regression Forests algorithm proposed by Dantone et al [35], the Random Forest Regression-Voting method described by Cootes et alor the Structured-Output Regression forest method [36]. In general, these approaches are more robust than the previous ones due to the combination of votes from different face regions, but they have not demonstrated a good balance between accuracy and efficiency for face alignment in-the-wild [1].…”