Table 4.
Comparison of different methods used for apple mealiness detection
| Method | Cultivar | Groups | Classifier | Accuracy (%) | Reference |
|---|---|---|---|---|---|
| Time resolved spectroscopy | Cox | Mealy and non-mealy | DA | 80 | Valero et al. (2005) |
| Ultrasonic | Cox | Fresh and overripe mealiness | LDA | 94.2 | Bechar et al. (2005) |
| Magnetic resonance imaging | Top-Red | Fresh, intermediate and mealy | SDA | 87.5 | Barreiro et al. (2000) |
| Hyperspectral imaging | Red Delicious | Mealy and non-mealy | PLS-DA | 86.7 | Huang and Lu (2010) |
| Biospeckle imaging | Red Delicious | Fresh and semi-mealy | ANN | 79.8 | Arefi et al. (2016) |
| Fluorescence | Jonagold | Mealy, mid-mealy and non-mealy | SOM | 90 | Moshou et al. (2003) |
| Impact response | Golden Delicious | Mealy and healthy | DA | 80 | Arana et al. (2004) |
| Impact response | Red Delicious | Mealy and non-mealy | CNN-AlexNet | 91.1 | This study |
| Impact response | Red Delicious | Mealy and non-mealy | CNN-VGG19 | 86.9 | This study |
DA discriminant analysis, LDA linear discriminant analysis, SDA stepwise discriminant analysis, PLS-DA partial least squares discriminant analysis, ANN artificial neural networks, SOM self-organizing map, CNN convolutional neural networks