a convolutional neural network based automated method for detecting paddy leaf disease

India's main crop for production is paddy. Over the past eleven years, the agricultural sector has contributed around 18.08 % of India's GDP. Unfortunately, the farmers who put forth a lot of effort to raise this crop have to deal with significant losses due to crop damage brought on by many paddy illnesses. About seven or eight of the more than thirty paddy leaf diseases that exist are fairly prevalent in India. Among the various paddy leaf diseases, the most common and detrimental ones are Brown Spot Disease, Blast Disease, Bacterial Leaf Blight, and others. Paddy plant growth and productivity are being hampered by these diseases, which can result in significant financial and environmental losses. Crop damage can be significantly decreased and farmer losses can be avoided if these discases can be identified carly on with high precision and speed. Four disease categories and one paddy leaf class that is healthy have been covered in this paper. The primary goal of this paper is to provide the best results for paddy leaf disease detection by using deep learning CNN models for automated detection, which can achieve the highest accuracy, as opposed to the time- consuming manual disease detection process, which is also of questionable accuracy. After analyzing four models (VGG-19, Inception-Resnet-V2, ResNet-101, and Xception), it was discovered that Inception-Resnet-V2 had a greater accuracy of92.68%. Keywords: Paddy leaf disease, deep convolutional neural network (DNN), transfer learning, VGG-19, ResNet-101, Inception-ResNet-V2