Development of a flexible Computer Vision System for marbling classification

Abstract

Traditional marbling meat evaluation is a tedious, repetitive, costly and time-consuming task performed by panellists. Alternatively, we have Computer Vision Systems (CVS) to mitigate these problems. However, most of CVS are restricted to specific environments, configurations or muscle types, and marbling scores are settled for a particular marbling meat standard. In this context, we developed a CVS for meat marbling grading, which is flexible to different muscle colour contrasts and grading standards. Essentially, the proposed method segments an image pre-processed by illumination normalisation and contrast enhancement, analyses visible intramuscular fat pixels and attributes a score based on a desired meat standard defined in the learning step. Learning approach is an instance-based system making use of k-Nearest Neighbours algorithm (k-NN) to attribute a score from segmentation results. The algorithm classifies the new samples based on scores assigned by panellists. We investigated the optimal number of samples for modelling, focusing on the smallest number leading to acceptable accuracy, and considering two different animal species: bovine and 1 swine. The CVS led to accuracy values equal to 81.59% (bovine) and to 76.14% (swine), using only three samples for each marbling score.

Computer Vision Classification of Barley Flour Based on Spatial Pyramid Partition Ensemble

“…Barbon et al. [23] proposed a CVS for meat classification based on image features, managed by an instance-based system using k-NN to classify meat according to marbling scores from image features. The authors presented an accuracy of 81.59% for bovine and 76.14% for swine samples, using only three samples for each marbling score by the k-NN prediction models.…”

“…The proposed SPPe was evaluated in a CVS with a set of image features based on color, intensity, and texture, in comparison to SPP [21], directly using the features extracted from the Region Of Interest (ROI), as traditional CVS [13,22,23,24]. We compared the performance of four different machine learning algorithms: Random Forest (RF), Support Vector Machine (SVM), k Nearest Neighbor (k-NN), and J48 decision tree for modeling the classifier.…”

A Review on Meat Quality Evaluation Methods Based on Non-Destructive Computer Vision and Artificial Intelligence Technologies

“…And research shows that meat quality is the most important purchase parameter affecting a consumer's decision (Kamruzzaman et al, 2016a;Barbon et al, 2017).…”

“…With meat consumption growing, quality is becoming more and more important to consumer’s purchase decision ( Wei et al, 2019 ). And research shows that meat quality is the most important purchase parameter affecting a consumer’s decision ( Barbon et al, 2017 ; Kamruzzaman et al, 2016a ).…”

Digital Image Filtering Optimization Supporting Iberian Ham Quality Prediction

“…The combination of digital food image analysis (DFIA) and predictive techniques to assess meat quality is arousing interest [27] because it offers some advantages over the sensory analysis, such as being non-destructive, less time-consuming, and low cost [28]. The first steps that have been taken to predict pork meat by applying data mining (Iberian ham [29,30] and loin [21,31,32]) and machine learning (marbling [33,34]) are recent. Pork meat quality prediction improves when the fractal analysis framework is taken into account [21,32].…”