Automatic Early Diagnosis of Dome Galls in Cordia Dichotoma G. Forst. Using Deep Transfer Learning

There are several infections and diseases that can affect plants. Diagnosis of plant is a challenging task. A fascinating method for identifying computer-based diagnosis using digital images plant’s leaves. Most the earlier research in this field has been devoted to feature engineering traditional machine learning (ML) techniques. Based on hand-crafted features taken from photographs leaves, these methods identify various diseases. It be use extract high-quality images. Deep (DL) algorithms, have relieved by automatically extracting resilient features. However, high number parameters conventional DL models typically leads overfitting. Therefore, gradient vanishing problems vast networks intensify failure generalization errors. Additionally, getting big dataset starting scratch train deep model also difficult. To solve issues detect Dome Galls Cordia dichotoma G. Forst. early, suggests quick efficient transfer (DTL) model. In proposed method, custom leaf created. 1784 collected real world environment, offline augmentation techniques applied obtaining final training 5400 Further, image preprocessing used enhance model, Yolov4, modified trained early detection dome galls leaves Cordia. The pre-trained weights, process called (TL). tested another set 200 images, results show an accuracy 95% F1-score 95.8%. Experimental Yolov4 performs 3% more accurately than original Yolov4.