Abstract
In a daily situation, when we try to buy a vegetable, we, first of all, rely on visual impression of the object for the first evaluation, and then use tactile information for a final decision-making. Our aim of this study is to investigate what of visual information is extracted from vegetable images, in which their degradation progresses gradually in time, for the estimation of the freshness, and to clarify how such visual information contributes. We tested three different kinds of vegetable images. Our ten observers (9 male and one female undergraduates) visually evaluated the freshness of 76 images with seven grades. In these images, the freshness of the vegetables was gradually degraded every hour. To find a dominant factor for the evaluation, we applied PCA by using a mean chromaticity difference in the u'v' color space, a mean luminance, and size of a vegetable in the image as the data set. In the result, the visual evaluation of the freshness of the vegetable gradually decreased, and the first principle component of the data set monotonically lowered as well. This indicates that the first principle component can represent the visual evaluation as for the freshness of the vegetables with highly correlation.
Meeting abstract presented at OSA Fall Vision 2012