Simhi Noa, Yovel Galit
The Sagol School of Neuroscience, Tel-Aviv University, Tel-Aviv 69978, Israel; The School of Psychological Sciences, Tel-Aviv University, Tel-Aviv 69978, Israel.
Vision Res. 2016 May;122:12-20. doi: 10.1016/j.visres.2016.02.003. Epub 2016 Mar 24.
While the importance of faces in person recognition has been the subject of many studies, there are relatively few studies examining recognition of the whole person in motion even though this most closely resembles daily experience. Most studies examining the whole body in motion use point light displays, which have many advantages but are impoverished and unnatural compared to real life. To determine which factors are used when recognizing the whole person in motion we conducted two experiments using naturalistic videos. In Experiment 1 we used a matching task in which the first stimulus in each pair could either be a video or multiple still images from a video of the full body. The second stimulus, on which person recognition was performed, could be an image of either the full body or face alone. We found that the body contributed to person recognition beyond the face, but only after exposure to motion. Since person recognition was performed on still images, the contribution of motion to person recognition was mediated by form-from-motion processes. To assess whether dynamic identity signatures may also contribute to person recognition, in Experiment 2 we presented people in motion and examined person recognition from videos compared to still images. Results show that dynamic identity signatures did not contribute to person recognition beyond form-from-motion processes. We conclude that the face, body and form-from-motion processes all appear to play a role in unfamiliar person recognition, suggesting the importance of considering the whole body and motion when examining person perception.
虽然面孔在人脸识别中的重要性已成为众多研究的主题，但相对较少有研究考察动态中的整个人体识别，即便这与日常体验最为相似。大多数考察动态中人体的研究使用点光显示，这种显示有诸多优点，但与现实生活相比显得匮乏且不自然。为了确定在识别动态中的整个人体时使用了哪些因素，我们使用自然主义视频进行了两项实验。在实验1中，我们采用了匹配任务，其中每对中的第一个刺激可以是一段视频或来自全身视频的多幅静态图像。进行人脸识别的第二个刺激可以是全身图像或仅面部图像。我们发现，身体对人脸识别的贡献超过了面部，但这仅在接触动态之后。由于人脸识别是在静态图像上进行的，动态对人脸识别的贡献是由动态形状感知过程介导的。为了评估动态身份特征是否也有助于人脸识别，在实验2中，我们展示了动态中的人物，并将视频中的人脸识别与静态图像中的人脸识别进行了比较。结果表明，动态身份特征对人脸识别的贡献并不超过动态形状感知过程。我们得出结论，面部、身体和动态形状感知过程似乎都在陌生人识别中发挥作用，这表明在考察人际感知时考虑全身和动态的重要性。
