Authors: Christopher Kucha, Li Liu, Michael Ngadi, Claude Gari?py
Identifier: CSBE19116
Download file:
Published in: CSBE-SCGAB Technical Conferences » AGM Vancouver 2019

Download RAW file:
DOI:
Description: The amount of intramuscular fat (IMF) in the meat is associated with the juiciness, flavor, tenderness and nutritional value. Therefore, the quantification of IMF is essential for assessing the overall consumer acceptability of the meat. The measurement of (IMF) currently relies on extraction techniques. These methods are labor-intensive and time-consuming and therefore require a rapid alternative. Hyperspectral imaging (HSI) is one such technique that is gaining popularity in the meat industry. In the use of this technology, the spectral information, as well as the imaging information are obtained which reveals the spatially-resolved spectral properties of the material. Traditionally HSI primarily focuses on the spectral information with minimal utilization of the spatial information present in the data. Consequently, this investigation aimed to simultaneously utilize the spectral and spatial information to achieve an efficient quantification of the IMF in pork loin cutlets. The textural information properties were extracted using statistical properties of the gray level co-occurrence matrix (GLCM). The textural properties were combined with the spectral information using one of the feature-level data fusion. Partial least square was used to build the regression models from the fused data. The result showed that the fusion of textural and spectral information resulted in an efficient quantification of IMF content in pork than the use of either spectral or textural information in isolation.

Keywords: Hyperspectral imaging, pork loin, intramuscular fat, data fusion
Résumé:
Mots-clés:
Citation:
Volume:
Issue:
Pages -
Contributor:
Date: 2019-07-15
Technical field:
Conference name: CSBE/SCGAB 2019 Annual Conference, Vancouver, BC, 14-17 July 2019.
Session name: Electronic and Instrumentation 2

Other information:
Type: Presentation
Format:
Publication type: Text.Abstract
Source:
Relation:
Coverage: Canada
Language 1: en
Language 2:
Rights: Canadian Society for Bioengineering
Notes:
Other files: