Analysis of food appearance properties by computer vision applying ellipsoids to colour data
The use of computer vision for estimating composition in food products has become wide spread in recent years, especially for products where by measuring colour or other spectral features, we are able to estimate the composition, variety, or ripeness. On the other hand, the appearance is one of the...
| Autores: | , , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión aceptada para publicación |
| Fecha de publicación: | 2013 |
| País: | España |
| Institución: | Universidad de Sevilla (US) |
| Repositorio: | idUS. Depósito de Investigación de la Universidad de Sevilla |
| OAI Identifier: | oai:idus.us.es:11441/155853 |
| Acceso en línea: | https://hdl.handle.net/11441/155853 https://doi.org/10.1016/j.compag.2013.08.027 |
| Access Level: | acceso abierto |
| Palabra clave: | CIELAB Clustering Colour ellipsoids Food colour Image analysis |
| Sumario: | The use of computer vision for estimating composition in food products has become wide spread in recent years, especially for products where by measuring colour or other spectral features, we are able to estimate the composition, variety, or ripeness. On the other hand, the appearance is one of the most worrying issues for producers due to its influence on quality and consumer preferences. Computer vision is the best option to satisfy the need of measuring colour and heterogeneity in these products. Previous studies have expressed the heterogeneity with the standard deviation or other magnitudes that do not explain accurately the distribution of colour in CIELAB colour space. Graphing the scatterplot of the a*b* values belonging to the pixels of the image helps to explain the shape of the point cloud, but currently there is not an objective procedure to quantify these point clouds. This work has established a method for improving the illustration of colour measurements by image analysis. The proposed algorithm classified the point clouds by clustering methods and established the methodology for fitting the resulting clusters into ellipsoids. Their geometric features described the shape of the clouds in a quantitatively manner and they could be useful in multivariate statistical techniques for classifying and predicting chemical properties. |
|---|