臉孔辨識歷程為個體如何從收錄、儲存進而提取臉孔的認知運作過程。根據神經心理學、臉孔失認症病人及臉孔倒立效果的研究,臉孔辨識確實與物體及文字辨識有所不同之處。雖然以往的臉孔辨識模式,例如功能模式、互動激發模式及型態調適模式,皆試圖對個體如何辨識臉孔提出解釋,但是針對臉孔的表情訊息及熟悉度問題,似乎都未有完善的說明,因此本研究提出的二階段臉孔辨識模式主要目的即是希望驗證表情訊息及熟悉度對於臉孔辨識歷程之影響。 二階段臉孔辨識模式認為在臉孔收錄歷程中,表情分析和特徵整體分析雖然分離為兩條路徑,但是在產生臉孔內在表徵時,這兩類的訊息將加以整合收錄,因此臉孔內在表徵應包含表情訊息,而非平靜臉的儲存形式。此外,隨著熟悉度增加,個人的臉孔節點會越獨特且清晰,因此在提取高熟悉度臉孔時,會較自動化且不受臉孔呈現方式的影響。同時個體在儲存臉孔時,相似性越高的節點,彼此的距離越接近,因此容易產生競爭效果。實驗一利四種變化型態臉孔,證實臉孔的內在表徵包含表情訊息,而非平靜臉的儲存形式。實驗二利用三種不同的作業要求,證實表情和臉孔特徵訊息會整合至臉孔表徵中。實驗三利用呈現方式及集合數目的不同,證實熟悉度越高,臉孔節點越獨特清晰,但結果發現熟悉臉孔的提取仍受呈現方式及集合數目的影響。最後實驗四利用高低相似度的臉孔,證實相似度越高的臉孔節點越接近,也越容易產生競爭效果。 本研究的結果支持了表情訊息對於臉孔辨識歷程是一重要的因素,也證實熟悉性對臉孔儲存的影響。不過在臉孔的提取上,熟悉度效果仍需做進一步的驗證。對後續的研究方向,本研究亦提供了一些建議。
Face recognition is a cognitive processing for individual to encode、store and retrieve face. The studies from neuropsychology、prosopagnosia and face inversion effect suggested that face recognition processing can be indeed distincted from object and word recognition . Although previous models of face recognition (e.g., Functional Model、Interactive Activation Model and Pattern Adjust Model)tried to clarify the processes of face recognition, but there were two problems remained:the roles of emotional clue and familiarity. Therefore a two-stage face recognition model was proposed account for the effect of emotional clue and familiarity on face recognition. Two-stage face recognition model took the emotion analyses and features-whole analyses as two separate perceptual routes, and these information will integrate into a internal facial representation. Therefore internal facial representation should include emotional information, rather than an expressionless face. On the other hand, as the familiarity increased, individual face node would become clear, or even unique. It helps individual to retrieve face more automation easily and stable. While individual stored familiar faces, the distance between nodes would be too close to induce a competition effect. Therefore, four variation pattern faces were used in study 1 to verify face internal representation is not an expressionless one. Study 2 uses three different tasks to verify emotion and features-whole information will integrate into a internal facial representation. Study 3 uses two display ways and two collection sizes to verify that face nodes would be unique and clear as familiarity increased. But we find that familiar face was still influenced by display way and collection size. Finally, study 4 uses high/low familiar faces to verify the role of similarity on whether it could shorten the distance between face nodes and causes competition effect. This study demonstrates emotion clues and familiarity play important roles in face recognition. More studies are required to clarify the effect of familiarity. And some suggestions for the future study were also presented.