Purchase this article with an account.
Eiji Kimura, Yusuke Takano; Efficient brightness averaging of heterogeneous achromatic patches. Journal of Vision 2015;15(12):631. doi: 10.1167/15.12.631.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Mean brightness in a variegated region may work as a clue to illumination intensity over the region and play an important role in the perception of object lightness. This study investigated whether brightness can be efficiently averaged for heterogeneous achromatic patches. Experiment 1 investigated discrimination thresholds for mean brightness between two arrays of 12 heterogeneous patches of different luminances. The thresholds were compared to brightness discrimination thresholds between two arrays of 12 homogeneous patches and to those between two single patches. The two arrays (or patches) were simultaneously presented for 200 msec and followed by a pattern mask. Results showed that mean brightness judgments for heterogeneous arrays were as accurate as simple brightness comparison for single patches, although they were slightly worse than brightness judgments for homogeneous arrays. This finding is consistent with efficient brightness averaging of different luminance patches. However, additional experiments revealed that inexperienced naive observers may use shortcuts for mean brightness judgments; they tended to choose as the brighter array the one containing a highest luminance patch or the one consisting of the larger number of patches. To investigate the effects of these confounding factors, Experiment 2 measured discrimination thresholds for mean brightness between two arrays composed of different numbers of heterogeneous patches (6 vs. 12 or 9 vs. 12). The highest luminance patch was included in the array consisting of either the smaller or the larger number of patches, and thus using this clue for mean judgments would lead to highly biased thresholds. Results were consistent with brightness averaging, but a small bias (varying in the magnitude among observers) was found to choose the array containing the highest luminance patch. Overall, the present findings suggest that brightness can be efficiently averaged, but with a greater weight to the highest luminance.
Meeting abstract presented at VSS 2015
This PDF is available to Subscribers Only