Context-Aware Face Detection for Occluded Faces

Detecting of occluded faces is a real challenge for face detection in the computer vision area. This challenge tends to be higher when the occlusion is more than half of the face. Regardless to the great success of recent face detection systems, face detection with occlusion is not yet saturated and requires more attention. We proposed in this work an occluded faces dataset in high degree of occlusion along with context-based labelling, which is a useful technique that let CNN to learn more features during training. The experimental benchmark result showed weak performance of current face detection models in our proposed dataset.