Invited Talk at TSoT 21

Color distribution of fresh and ripe fruits

The visual quality assessment of food is guided by our visual perception and food interaction experience. We choose the food to purchase or consume by subjectively judging their physicochemical properties based primarily on their visual characteristics. In particular, for natural textures, such as the skin of fleshy fruits, we relate their taste to subjective quality assessments, such as its freshness and ripeness perception. The concept of freshness relates to the perception of degradation in the development cycle of a natural product. For example, a banana has a higher freshness when its skin is green; then, dark spots will cover its surface till all surface is black, at the lowest freshness stage. However, freshness must not be confused with ripeness. Ripeness perception is a particular stage of the fruit development cycle associated with the optimum moment for consumption. To understand the different visual cues that affect freshness and ripeness perception in fruits, we took pictures of four types of tropical fruits (avocado, banana, lulo, and guava) at different stages of their development cycle. Then, we presented calibrated pictures of these fruits to participants using a within-subject experimental design and a 2AFC paradigm where they chose, from a pair of images of fruits, which one was the oldest and which one they would eat. Experimental results revealed a high correlation between image luminance statistics and ripeness and freshness perception depending on the fruit. Moreover, color distribution was correlated with either freshness (mostly channel a) or ripeness (mostly channel b) perception. For some fruits, freshness perception was associated with a lower a-channel value (green) whereas, a high variance in b-channel (blue) was associated with high ripeness perception. However, there was no general model that fitted all our results supporting the idea that fruit visual quality is dependent on the particular fruit in question.

Check out the video part of the program at TSoT’21