Abstract
Aim: To investigate the role of shading and reflectance information in the recognition of real images of human faces. Methods: Pictures of human faces were taken with a calibrated digital camera and subjected to a reflectance-shading separation algorithm similar to that described by Olmos & Kingdom, Perception, 33, 2004. Subjects were presented with nine possible combinations of original, reflectance-only and shading-only faces in various forced-choice, test-versus-comparison combinations. Subjects were required to indicate on each trial whether the test and comparison faces were from the same face or not. Results: Removing either the reflectance or the shading components from an original face impaired recognition. Subjects found it more difficult to judge whether a reflectance-only and shading-only pair were from the same face, compared to pairs that included the original face. Subjects found it more difficult to compare two shading-only faces compared to either two reflectance-only, or two original faces. Conclusion: Although both shading and reflectance information appears to be important for face recognition, shading may be less important than previously thought.
Supported by Canadian Institute of Health Research grant #MOP-11554 given to F.K.
Supported by Canadian Institute of Health Research grant #MOP-11554 given to F.K.