Face Recognition with Multi-Tone and Two-Tone Photographic Negatives

Abstract

The effects of photographic negatives on face recognition are often studied in two different ways—faces are learned and tested in photographic negatives (NN) or the contrast is reversed between learning and testing, ie they are learned in positives and tested in negatives (PN) or vice versa (NP). We have examined recognition performance for faces under these three conditions along with a control condition where faces were learned and tested in positives (PP). Using multi-tone face images, we found that the effect of photographic negatives was more pronounced in PN and NP than in NN. No differences were found between PN and NP or between NN and PP. When two-tone face images were used, recognition performance was worse in all conditions, except PP, when compared to the multi-tone counterparts. Our results show that contrast incongruency between learning and testing is the predominant factor affecting performance and that deficits in sensory coding or retention of negative face images are unlikely to be major factors. The advantage of multi-tone over two-tone negatives can be attributed to preserved facial information carried by the high-spatial-frequency components of the image.