Surface color distribution analysis by computer vision compared to sensory testing: Vacuum fried fruits as a case study

Fitriyono Ayustaningwarno,Vincenzo Fogliano,Ruud Verkerk,Matthijs Dekker

doi:10.1016/j.foodres.2021.110230

Fitriyono Ayustaningwarno, Vincenzo Fogliano + Show 2 more

Open Access

https://doi.org/10.1016/j.foodres.2021.110230

Copy DOI

Abstract

Color is a main factor in the perception of food product quality. Food surfaces are often not homogenous at micro-, meso-, and macroscopic scales. This matrix can include a variety of colors that are subject to changes during food processing. These different colors can be analyzed to provides more information than the average color. The objective of this study was to compare color analysis techniques on their ability to differentiate samples, quantify heterogeneity, and flexibility. The included techniques are sensory testing, Hunterlab colorimeter, a commercial CVS (IRIS-Alphasoft), and the custom made CVS (Canon-CVS) in analyzing nine different vacuum fried fruits. Sensory testing was a straightforward method and able to describe color heterogeneity. However, the subjectivity of the panelist is a limitation. Hunterlab was easy and accurate to measure homogeneous samples with high differentiation, without the color distribution information. IRIS-Alphasoft was quick and easy for color distribution analysis, however the closed system is the limit. The Canon-CVS protocol was able to assess the color heterogeneity, able to discriminate samples and flexible. As a take home massage, objective color distribution analysis has a potential to unlock the limitation of traditional color analysis by providing more detailed color distribution information which is important with respect to overall product quality.

Highlights

Food color and color distribution, in raw and processed products, are one of the key sensory quality characteristics for consumers and deter mine the acceptation of a food product
Overall color analysis by sensory testing was done by selecting a color within the 920 colors of the Royal Horticultural Society (RHS) color chart which represents the overall color of the sample
Despite the panelists were selected to have an appropriate color discrimination ability, when they analyzed the fruit, each panelist had a different perception of the color of the samples, which lead to a variety of colors perceived among the panelists (Fig. 3A)

Summary

Introduction

Food color and color distribution, in raw and processed products, are one of the key sensory quality characteristics for consumers and deter mine the acceptation of a food product. Color assessment is highly relevant in the total quality assessment of food products (Krokida, Oreopoulou, Maroulis, & Marinos-Kouris, 2001). Efforts to measure color of foods have been done using a variety of methods with the final aim of product quality analysis and standardization. Color analysis by sensory testing has been done for quality control of food production. Visual evaluation of color via sensorial experiments is valuable as it can be expanded to the consumer’s demands and satisfaction. In general sensory testing has some limitations, including subjectivity, response variation and limitation in the number of samples (Borras et al, 2015)

Objectives

Methods

Results

Conclusion