基于事例的面部动作单元识别算法

本文提出基于事例的交互式遗传算法进行面部动作单元识别的算法,将用户的比较能力融入到搜索过程,快速检索到与待识别图像匹配的事例图像,从而实现动作单元的半自动识别。该方法不需抽取图像特征,因而可用于识别非控制成像条件下自发面部图像或图像序列中的动作单元,具有较好的鲁棒性和实用性。文中采用16幅受控成像条件下收集的简单图像进行实验,单独AU的平均识别率达到77.5%,AU组合的平均相似度为82.8%。采用10幅有干扰的非受控成像条件下收集的复杂图像进行实验,单独AU的平均识别率为82.8%,AU组合的平均相似度为93.1%。相对于特征脸算法,本文算法的平均识别率和相似度都有较大程度的提高。

Algorithm of Case-based Facial Action Units Recognition

This paper proposes a case-based automatic facial action units recognition approach using interactive genetic algorithm, which embeds human compare ability into search process. To obtain AU codes of a new facial image, IGA is applied to retrieve a match instance to this new one from the case base based on users~ evaluation. Then the solution suggested by the matching case is used as the AU codes to the new facial images. The approach is easy and has the latent capacity to detect AU codes from facial images and image sequences insensitive to pose and occlusions, since feature extraction is not needed. The effectiveness of our approach is evaluated by 16 standard images and 10 un-standard images. A recognition rate of 77.5% is achieved on single AUs, and a similarity rate of 82.8% is obtained on AU combinations to standard images. A recognition rate of 82.8% is achieved on single AUs, and a similarity rate of 93.1% is obtained on AU combinations to un-standard images. The compare experiments with eigenface method also demonstrate the effectiveness of our approach.