Login / Signup

Exploring the relationship of colour categorization with pork colour standards under different subjective and objective conditions.

B UttaroS ZawadskiM Juárez
Published in: Meat science (2024)
The impact of using incorrect lighting while subjectively scoring pork colour with subjective standards (Japanese, Canadian, and Kodak grey) was explored. Lightness was more important than a good colour match between standards and meat. Subjective and image-based automated scoring with Canadian standards were correlated at 0.71-0.77 (P < 0.001) with significant differences in scale distribution (D = 0.14-0.46; P < 0.002), primarily with moderately dark meat. Automated scoring on full colour and greyscale images were strongly related (r = 0.83, P < 0.001) and showed matching score distributions when whole scores were used. Tracking automated colour categorization during blooming showed very good potential for reliable categorization after 1 min exposure to air for most meat colours, indicating that reliable automated on-line sorting of pork for colour is easily within reach.
Keyphrases
  • deep learning
  • machine learning
  • high throughput
  • convolutional neural network
  • sleep quality
  • multiple sclerosis
  • depressive symptoms