Skip to main content

Advertisement

Table 2 Comparison of detection performance of our model (CNN) with our previous model [19] on the validation set

From: Automatic estimation of heading date of paddy rice using deep learning

Validation dataset No. of images Sliding window Precision Recall F1-score
Dimensions Stride [19] CNN [19] CNN [19] CNN
Kinmaze 15 \(140\, \times \,140\) 140 0.81 0.96 0.73 0.68 0.77 0.80
Kamenoo 15 \(160\, \times \,160\) 140 0.67 0.84 0.61 0.71 0.64 0.77
Koshihikari-1 15 \(160\, \times \,160\) 150 0.72 0.80 0.65 0.70 0.68 0.70
Koshihikari-2 15 \(160\, \times \,160\) 150 0.72 0.84 0.73 0.75 0.72 0.79
Koshihikari-3 15 \(160\, \times \,160\) 150 0.74 0.89 0.69 0.71 0.71 0.79