基于多尺度注意力机制相位展开的三维人脸建模  

Three-Dimensional Face Modeling Based on Multi-Scale Attention Phase Unwrapping

作　　者：朱江平 王睿珂 段智涓 黄怡洁 何国欢 周佩 Zhu Jiangping;Wang Ruike;Duan Zhijuan;Huang Yijie;He Guohuan;Zhou Pei(College of Computer Science,Sichuan University,Chengdu,Sichuan 610065,China;National Key Laboratory of Fundamental Science on Synthetic Vision,Sichuan University,Chengdu,Sichuan 610065,China)

机构地区：[1]四川大学计算机学院,四川成都610065 [2]四川大学视觉合成图形图像技术国防重点学科实验室,四川成都610065

出　　处：《光学学报》2022年第1期155-166,共12页Acta Optica Sinica

基　　金：国家自然科学基金(61901287,62101364);四川省重点研发专项(2021YFG0195,2020YFG0112,2020YFG0306);四川省重大科技专项(2019ZDZX0039,2018GZDZX0029)。

摘　　要：相位展开作为三维(3D)测量技术中的关键环节,其解析精度直接影响3D建模的精度。由于存在欠采样和相位不连续等问题,故传统空间相位展开难以得到正确的相位信息,而时间相位展开又需要额外的信息辅助。针对复杂场景中的3D人脸建模,提出了基于多尺度注意力机制的相位展开网络。在所提网络中,利用编码器-解码器结构融合多尺度特征,并在解码网络中嵌入注意力子网络以获取上下文信息。构建一个包含5000组样本的FACE数据集和一个包含100组样本的MASK数据集,每组样本均包含截断相位和连续相位的真值,这些真值可用于相位展开的训练及测试。所提网络在FACE数据集和MASK数据集上的均方根误差分别为0.0387 rad和0.0273 rad,结构相似性分别为0.9850和0.9793。在欠采样、相位不连续等区域中,所提网络可快速准确地提取相位特征,进而保证了相位展开的正确性。最后,通过对比实验证实了所提网络的有效性和可行性。Phase unwrapping plays an important role in three-dimensional(3 D) measurement technologies, and its analytical accuracy directly affects the accuracy of 3 D modeling. Due to undersampling and discontinuity of the wrapped phase, it is difficult to obtain correct phase information for traditional spatial phase unwrapping, while temporal phase unwrapping requires additional auxiliary information. For 3 D face modeling in complex scenarios, a phase unwrapping network based on multi-scale attention is proposed in this paper. In this network, the encoder-decoder structure is used to fuse multi-scale features, and an attention sub-network is embedded into the decoding network for contextual information collection. A FACE dataset of 5000 samples and a MASK dataset of 100 samples are constructed, and each sample contains the truth values of wrapped phases and continuous phases for training and testing of phase unwrapping. The root-mean-square errors of the proposed network are 0.0387 rad and 0.0273 rad on the FACE dataset and the MASK dataset. The structural similarities are 0.9850 and 0.9793 respectively. The phase features can be extracted quickly and accurately in areas such as undersampled and phase discontinuous ones to ensure the correctness of phase unwrapping. Finally, the effectiveness and feasibility of the proposed network are verified by comparative experiments.

关 键 词：测量 三维人脸建模 相位展开 多尺度注意力机制融合 上下文特征信息 编码器-解码器结构 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]