Color-Aware Two-Branch DCNN for Efficient Plant Disease Classification

Deep convolutional neural networks (DCNNs) have been successfully applied to plant disease detection. Unlike most existing studies, we propose feeding a DCNN CIE Lab instead of RGB color coordinates. We modified an Inception V3 architecture include one branch specific for achromatic data (L channel) and another chromatic (AB channels). This modification takes advantage the decoupling information. Besides, splitting branches reduces number trainable parameters computation load by up 50% original figures using layers. achieved state-of-the-art classification accuracy 99.48% on Plant Village dataset 76.91% Cropped-PlantDoc dataset.