摘要
This paper presents a new face detection approach to real-time applications, which is based on the skin color model and the morphological filtering. First the non-skin color pixels of the input image are removed based on the skin color model in the YC rC b chrominance space, from which we extract candidate human face regions. Then a mathematical morphological filter is used to remove noisy regions and fill the holes in the candidate skin color regions. We adopt the similarity between the human face features and the candidate face regions to locate the face regions in the original image. We have implemented the algorithm in our smart media system. The experiment results show that this system is effective in real-time applications.
This paper presents a new face detection approach to real-time applications, which is based on the skin color model and the morphological filtering. First the non-skin color pixels of the input image are removed based on the skin color model in the YC rC b chrominance space, from which we extract candidate human face regions. Then a mathematical morphological filter is used to remove noisy regions and fill the holes in the candidate skin color regions. We adopt the similarity between the human face features and the candidate face regions to locate the face regions in the original image. We have implemented the algorithm in our smart media system. The experiment results show that this system is effective in real-time applications.
基金
ThisprojectwassupportedinpartbyaHongKongResearchGrantsCouncil (RGC)ProjectNo .CityUHK #90 4 0 6 90 873
aStrategicDevelopmentGrant (SDG)ProjectNo.#70 10 0 2 3 873
anAppliedResearchGrantProjectNo.#96 40 0 0 2 873andRCMTProjectNo .#936 0 0 80 873
from
