Abstract
Objective: Some studies show legibility of characters are strongly affected by their stroke frequencies (Majaj, et. al, 2002; Zhung et. al, 2007). Takahashi and Oda (2014) showed the legibility was well explained by stroke frequency and observers' visual acuity for Chinese characters in a very wide range of stroke frequency using outlined(OFS) and solid font style(SFS). Their results also showed a systematic difference in legibility between OFS and SFS which could not be explained by stroke frequency and observers' acuity. The objective is to see whether the critical contrast component for character recognition (1-4 cycles per character(cpc) component) explains the difference. Method: Stimuli and apparatus were the same as Takahashi and Oda(2014). Stimuli were Chinese characters of 3 complexity levels, high: 16, middle: 8, and low: 2-3 stroke frequencies per character with 10 characters in each group. These 30 characters were presented in two font styles, i.e., OFS and SFS of Kozuka Gothic ProN. The line width of OFS was 1/24 of character height. Twelve university students with normal acuity participated in the threshold measurement and legibility was defined as character size for 50% recognition threshold for each character. The critical contrast component of 1-4 cpc was calculated with FFT for each character. We compared the coefficient of determination between two multiple regression analyses, i.e., one by stroke frequency and the other with stroke frequency and the critical contrast component. Result and Discussion: The R^2 for the model with two factors was 0.993 and better than one for the model with stroke frequency only. However, the resulted regression model predicts that characters with higher contrast components are less legible. We suspect that the effect of the critical contrast component would depend on stroke frequency.
Meeting abstract presented at VSS 2016