Fig. 3 | Scientific Reports

Fig. 3

From: Quality prediction of air-cured cigar tobacco leaf using region-based neural networks combined with visible and near-infrared hyperspectral imaging

Fig. 3The alternative text for this image may have been generated using AI.

Prediction results by three models. (a) moisture PLSR; (b) moisture CNN; (c) moisture DR-CNN; (d) chlorophyll PLSR; (e) chlorophyll CNN; (f) chlorophyll DR-CNN; (g) nitrogen PLSR; h) nitrogen CNN; (i) nitrogen DR-CNN; (j) sugar PLSR; (k) sugar CNN; (l) sugar DR-CNN. The red line in each panel represents the linear fit of the dataset. The closer the scatter points are to the line, the better the model’s predictive accuracy and the stronger the correlation between the predicted and actual values. A tighter distribution of points around the line indicates more stable model performance with lower variance. Conversely, a wider spread of points around the fitting line suggests higher variance and less accurate predictions.

Back to article page