Improving Unfamiliar Face Matching Using Multiple Images and Feedback
Improving Unfamiliar Face Matching Using Multiple Images and Feedback
复制标题
使用多个图像和反馈改进陌生的面部匹配
作者：
Dowsett A
众所周知，我们匹配陌生面孔的能力很容易出错（Burton & Jenkins，2011）。尽管如此，我们仍然严重依赖照片识别来证明我们的身份。相反，当匹配熟悉的面孔时，我们的表现几乎完美（Bruce et al., 2001）。当我们熟悉一张脸时，我们能够克服人体内固有的变异性来进行准确的匹配（Jenkins et al, 2011）。然而，当我们不熟悉时，我们缺乏这种经验，被迫进行基于图像的比赛。 在这一系列的研究中，我们询问这种人体内的变异性是否可以用来训练个体进行不熟悉的面部识别。我们使用一项困难的面部匹配任务来表明，人体内变异性的存在可以提高目标个体的表现，但通常不能提高面部匹配的表现。此外，还引入了逐个试验的反馈来证明性能可以进一步改进。我们考虑这些结果对训练面部匹配性能的影响。
It is well established that our ability to match unfamiliar faces is highly error prone (Burton & Jenkins, 2011). Yet despite this we still rely heavily on photographic identification to prove our identity. Conversely, when matching familiar faces, we perform almost perfectly (Bruce et al., 2001). When we are familiar with a face we are able to overcome inherent within-person variability to make an accurate match (Jenkins et al, 2011). However, when we are unfamiliar, we lack this experience and are forced into a pictorial based match. In this series of studies we ask whether this within-person variability can be used to train individuals in unfamiliar face recognition. We used a difficult face matching task to show that the presence of within-person variability can improve performance for target individuals, but not for face matching in general. Furthermore, trial-by-trial feedback was introduced to demonstrate performance can be improved upon further. We consider the implications of these results for training face matching performance.