Caplova Zuzana, Compassi Valentina, Giancola Silvio, Gibelli Daniele M, Obertová Zuzana, Poppa Pasquale, Sala Remo, Sforza Chiarella, Cattaneo Cristina
Università degli Studi di Milano, Dipartimento di Scienze Biomediche per la Salute, Via Mangiagalli 31, 20133 Milano, Italy.
Università degli Studi di Milano, Dipartimento di Scienze Biomediche per la Salute, Via Mangiagalli 31, 20133 Milano, Italy.
Sci Justice. 2017 Jul;57(4):250-256. doi: 10.1016/j.scijus.2017.03.005. Epub 2017 Mar 31.
The situation of missing children is one of the most emotional social issues worldwide. The search for and identification of missing children is often hampered, among others, by the fact that the facial morphology of long-term missing children changes as they grow. Nowadays, the wide coverage by surveillance systems potentially provides image material for comparisons with images of missing children that may facilitate identification. The aim of study was to identify whether facial features are stable in time and can be utilized for facial recognition by comparing facial images of children at different ages as well as to test the possible use of moles in recognition. The study was divided into two phases (1) morphological classification of facial features using an Anthropological Atlas; (2) algorithm developed in MATLAB® R2014b for assessing the use of moles as age-stable features. The assessment of facial features by Anthropological Atlases showed high mismatch percentages among observers. On average, the mismatch percentages were lower for features describing shape than for those describing size. The nose tip cleft and the chin dimple showed the best agreement between observers regarding both categorization and stability over time. Using the position of moles as a reference point for recognition of the same person on age-different images seems to be a useful method in terms of objectivity and it can be concluded that moles represent age-stable facial features that may be considered for preliminary recognition.
失踪儿童的情况是全球最牵动人心的社会问题之一。寻找和识别失踪儿童往往受到诸多因素的阻碍，其中包括长期失踪儿童的面部形态会随着成长而发生变化。如今，监控系统的广泛覆盖有可能提供图像资料，用于与失踪儿童的图像进行比对，从而便于识别。本研究的目的是通过比较不同年龄段儿童的面部图像，确定面部特征在时间上是否稳定以及能否用于面部识别，同时测试痣在识别中的可能用途。该研究分为两个阶段：（1）使用《人类学图谱》对面部特征进行形态学分类；（2）在MATLAB® R2014b中开发算法，以评估痣作为年龄稳定特征的用途。通过《人类学图谱》对面部特征进行评估显示，观察者之间的不匹配率很高。平均而言，描述形状的特征的不匹配率低于描述大小的特征。鼻尖裂和下巴酒窝在观察者之间关于分类和随时间的稳定性方面表现出最佳的一致性。将痣的位置作为在不同年龄图像上识别同一个人的参考点，从客观性角度来看似乎是一种有用的方法，可以得出结论，痣代表了年龄稳定的面部特征，可用于初步识别。
