Improved Search Strategies with Application to Estimating Facial Blendshape Parameters
This addresses a specific problem in computer vision for facial animation, where overfitting and underfitting hinder accurate expression interpretation, but it is incremental as it builds on existing optimization techniques.
The paper tackles overfitting in optimization for facial blendshape parameter estimation by proposing a method that reduces overfitting without regularization, which can cause underfitting, and demonstrates its effectiveness on 3D facial expression estimation problems.
It is well known that popular optimization techniques can lead to overfitting or even a lack of convergence altogether; thus, practitioners often utilize ad hoc regularization terms added to the energy functional. When carefully crafted, these regularizations can produce compelling results. However, regularization changes both the energy landscape and the solution to the optimization problem, which can result in underfitting. Surprisingly, many practitioners both add regularization and claim that their model lacks the expressivity to fit the data. Motivated by a geometric interpretation of the linearized search space, we propose an approach that ameliorates overfitting without the need for regularization terms that restrict the expressiveness of the underlying model. We illustrate the efficacy of our approach on minimization problems related to three-dimensional facial expression estimation where overfitting clouds semantic understanding and regularization may lead to underfitting that misses or misinterprets subtle expressions.