Toward Personalized Modeling: Incremental and Ensemble Alignment for Sequential Faces in the Wild
Toward Personalized Modeling: Incremental and Ensemble Alignment for Sequential Faces in the Wild
复制标题
DOI：
10.1007/s11263-017-0996-8
复制
发表时间：
2017-02
影响因子：
19.5
通讯作者：
Xi Peng;Shaoting Zhang;Yang Yu;Dimitris N. Metaxas
中科院分区：
文献类型：
--
作者：
Xi Peng;Shaoting Zhang;Yang Yu;Dimitris N. Metaxas
在无约束的视频上拟合面部标志是一项具有广泛应用的具有挑战性的任务。一般和联合对准方法都已提出，并取得了不同程度的成功。然而，许多通用方法对初始化非常敏感，并且通常依赖于离线训练的静态模型，这限制了它们在具有广泛变化的序列图像上的性能。另一方面，联合方法仅限于离线应用，因为它们需要所有帧进行批量对齐。为了解决这些限制，我们建议利用增量学习进行个性化的集成对齐。我们对多个初始形状进行采样，以在一帧内实现图像凝固，这使我们能够通过组稀疏正则化秩最小化来增量地进行整体对齐。同时，增量子空间自适应执行，以实现个性化建模在一个统一的框架。为了缓解漂移问题，我们利用一个非常有效的拟合评估网络来挑选出对齐良好的人脸，以进行鲁棒的增量学习。在受控和无约束数据集上进行的大量实验从不同方面验证了该方法的有效性，并在拟合精度和效率方面证明了其与现有技术相比的上级性能。
Fitting facial landmarks on unconstrained videos is a challenging task with broad applications. Bothgenericandjointalignment methods have been proposed with varying degrees of success. However, many generic methods are heavily sensitive to initializations and usually rely on offline-trained static models, which limit their performance on sequential images with extensive variations. On the other hand, joint methods are restricted to offline applications, since they require all frames to conduct batch alignment. To address these limitations, we propose to exploit incremental learning for personalized ensemble alignment. We sample multiple initial shapes to achieve image congealing within one frame, which enables us to incrementally conduct ensemble alignment by group-sparse regularized rank minimization. At the same time, incremental subspace adaptation is performed to achieve personalized modeling in a unified framework. To alleviate the drifting issue, we leverage a very efficient fitting evaluation network to pick out well-aligned faces for robust incremental learning. Extensive experiments on both controlled and unconstrained datasets have validated our approach in different aspects and demonstrated its superior performance compared with state of the arts in terms of fitting accuracy and efficiency.