ORFormer: Occlusion-Robust Transformer for Accurate Facial Landmark Detection
This addresses the issue of robust facial landmark detection under occlusions for applications in computer vision, though it is an incremental improvement over existing methods.
The paper tackled the problem of performance drops in facial landmark detection on partially occluded faces by introducing ORFormer, a transformer-based method that identifies and recovers features in non-visible regions, resulting in favorable performance against state-of-the-art methods on datasets like WFLW and COFW.
Although facial landmark detection (FLD) has gained significant progress, existing FLD methods still suffer from performance drops on partially non-visible faces, such as faces with occlusions or under extreme lighting conditions or poses. To address this issue, we introduce ORFormer, a novel transformer-based method that can detect non-visible regions and recover their missing features from visible parts. Specifically, ORFormer associates each image patch token with one additional learnable token called the messenger token. The messenger token aggregates features from all but its patch. This way, the consensus between a patch and other patches can be assessed by referring to the similarity between its regular and messenger embeddings, enabling non-visible region identification. Our method then recovers occluded patches with features aggregated by the messenger tokens. Leveraging the recovered features, ORFormer compiles high-quality heatmaps for the downstream FLD task. Extensive experiments show that our method generates heatmaps resilient to partial occlusions. By integrating the resultant heatmaps into existing FLD methods, our method performs favorably against the state of the arts on challenging datasets such as WFLW and COFW.