Two-Stage Intelligent DarkNet-SqueezeNet Architecture-Based Framework for Multiclass Rice Grain Variety Identification

Table 5

Classification results of fusion data (experiment 3).

Classifier (%)	Recall (%)	Sensitive (%)	F1 score (%)	FNR (%)	Accuracy (%)	Time (sec)
Linear SVM	99.96	99.94	99.95	0.04	99.9	275.3
Quadratic SVM	99.96	99.96	99.96	0.04	100	221.12
Weight KNN	99.88	99.7	99.79	0.12	99.9	133.5
Cosine KNN	99.86	99.88	99.87	0.14	99.9	146.3
Linear discriminant	99.76	99.76	99.76	0.24	99.7	143.63
Medium neural network	99.94	99.94	99.94	0.06	99.9	223.75
Narrow neural network	99.92	99.96	99.94	0.08	99.9	209.29
Wide neural network	99.96	99.96	99.96	0.04	100	139.86
Bilayered neural network	99.94	99.96	99.95	0.06	99.9	204.31
Trilayered neural network	99.9	99.92	99.91	0.1	99.9	229.86