Face Detection under Variable Lighting Based on Nine Point of Light

Different environment illumination has a great impact on face detection and recognition. In this paper, we present a solution based on the nine points of light. The basic idea is that there exists a configuration of nine points light source directions which can be acquired by taking nine images of each individual under these single sources. Using this set of nine directions, we construct a linear subspace for each collected example by rendering it under these different lighting conditions. And then by sampling several examples randomly from the linear subspace, the collected example set can be multiplied. The multiplied sample set is used to train a Support Vector Machine (SVM), which is tested on a test set. It turns out that the resulting subspace is effective at detection under a wide range of lighting conditions. To verify the generalization capability of the proposed method, we also use the expanded database to train an AdaBoost-based face detector and test it on the MIT+CMU frontal face test set. The experimental results also show that the data collection can be efficiently speeded up by the proposed methods.