Learning to Validate the Quality of Detected Landmarks
This addresses the problem of unreliable landmark detection in computer vision applications, particularly for facial analysis, though it appears incremental as it builds on existing regression-based methods.
The authors tackled the problem of validating the quality of detected image landmarks by introducing a new loss function that allows CNNs to estimate their own landmark accuracy, enabling exclusion of unreliable landmarks. They achieved this with a novel batch balancing approach that weights samples based on loss, tested on facial landmark datasets (300W, AFLW, WFLW).
We present a new loss function for the validation of image landmarks detected via Convolutional Neural Networks (CNN). The network learns to estimate how accurate its landmark estimation is. This loss function is applicable to all regression-based location estimations and allows the exclusion of unreliable landmarks from further processing. In addition, we formulate a novel batch balancing approach which weights the importance of samples based on their produced loss. This is done by computing a probability distribution mapping on an interval from which samples can be selected using a uniform random selection scheme. We conducted experiments on the 300W, AFLW, and WFLW facial landmark datasets. In the first experiments, the influence of our batch balancing approach is evaluated by comparing it against uniform sampling. In addition, we evaluated the impact of the validation loss on the landmark accuracy based on uniform sampling. The last experiments evaluate the correlation of the validation signal with the landmark accuracy. All experiments were performed for all three datasets.