A Novel Transient Wrinkle Detection Algorithm and Its Application for Expression Synthesis

Weicheng Xie Linlin Shen Jianmin Jiang

The ComputerVision Institute, School of Computer Science and Software Engineering, and the Shenzhen Key Laboratory of Spatial Information Smart Sensing and Services, Shenzhen University, Shenzhen 518060, China
the School of Computer Science, University of Nottingham, Nottingham NG8 1BB, U.K.
the Research Institute for FutureMedia Computing, Shenzhen University, Shenzhen 518060, China

Abstract

Because facial wrinkle is a representative feature of facial expression, automatic wrinkle detection has been an important and challenging topic for expression simulation, recognition, and animation. Recently, most works about wrinkle detection have focused on permanent wrinkles (e.g., age wrinkles), which are usually linear shapes, whereas the detection of transient wrinkles (e.g., expression wrinkles) has not been sufficiently studied because of their shape diversity and complexity. In this work, a novel algorithm for automatic detection of transient wrinkles with linear, fixed, and chaotic shapes is proposed, which largely consists of edge pair matching, active-appearance-modelbased wrinkle structure location, and support-vector-machinebased wrinkle classification. The proposed wrinkle detector is applied for expression synthesis and an improved Poisson wrinkle mapping approach is proposed. Experimental results illustrate the competitiveness of the proposed wrinkle detector in detecting different transient wrinkles. Compared with stateof- the-art algorithms, the proposed approach yields complete and accurate wrinkle centers. The expression synthesized by the improved wrinkle mapping is also much more realistic.

Visualization

Overview of the work

Framework of transient wrinkle detection

Illustration of wrinkle edge pair matching

(a) and (b) are two example faces from wrinkle database. (c) shows two fixed regions for AAM training and learning.

Illustration of wrinkle structure adjustment

Illustration of non-wrinkle region generation

Comparison of generated wrinkles with the initial (a) and improved (b) Poisson wrinkle mappings.

Detected wrinkles on the forehead regions

Forehead wrinkle location for 7th–11th frames of temporal expression sequence and the corresponding quantitative metric values.

Forehead wrinkle location on older adults with different conditions

Comparison of different operators for locating blurry wrinkle centerlines

Location results of threewrinkle detectors for the nasolabial wrinkles

Wrinkles verified by SVM

Detection results of six categories of wrinkles with relatively fixed shape

First row demonstrates the chin regions labeled with green polygons. The preceding and subsequent three images of the 2nd row present the correctly detected wrinkle and non-wrinkle regions, respectively.

Quantitative metric values for chin wrinkle and non-wrinkle regions

Wrong classification of chaotic chin wrinkles

Hybrid of chaotic and linear shape wrinkle detectors

Smile, laugh, and sad expressions synthesized by the proposed algorithm

Comparison of three wrinkle mapping algorithms for expression synthesis


@article{xie2016novel,
    title={A novel transient wrinkle detection algorithm and its application for expression synthesis},
    author={Xie, Weicheng and Shen, Linlin and Jiang, Jianmin},
    journal={IEEE Transactions on Multimedia},
    volume={19},
    number={2},
    pages={279--292},
    year={2017},
    publisher={IEEE}
}