ABSTRACT

Human unfamiliar face matching is error-prone, but some research suggests matching to multiple-image arrays instead of single images may yield improvements. Here, high or low variability arrays containing one, two, and three images, and a target image from the high and low variability image sets were displayed. Arrays were presented simultaneously or sequentially, and the target image was presented simultaneously with arrays or sequentially after arrays, in three experiments. Benefits from exposure to multiple images of the same person required simultaneous viewing of images and improvements were observed in match trials only. Only sequential viewing of a multiple-image array followed by a high variability target image enhanced overall accuracy across trial types, particularly for high variability arrays. Accuracy was highest when the target image and array items were visually similar. Results show the importance of image similarity, and suggest variability is most helpful when array and target are presented sequentially.

摘要

陌生的人脸匹配容易出错，但是一些研究表明，与多图像阵列而不是单图像匹配可能会带来改善。在此，显示包含一幅，两幅和三幅图像的高或低变异性数组，以及来自高和低变异性图像集的目标图像。在三个实验中，同时或依次显示阵列，并同时或依次在阵列后显示目标图像。暴露于同一个人的多张图像所带来的好处需要同时观看图像，并且仅在比赛试验中观察到了改善。仅顺序查看多图像阵列，然后是高可变性目标图像，可以提高整个试验类型的总体准确性，尤其是对于高可变性阵列。当目标图像和阵列项目在视觉上相似时，准确性最高。结果显示了图像相似性的重要性，并表明当顺序显示阵列和靶标时可变性最有帮助。